1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/ffmpeg_smp/benchmark.sh	Mon Aug 27 12:09:56 2012 +0200
     1.3 @@ -0,0 +1,126 @@
     1.4 +#! /bin/bash
     1.5 +
     1.6 +workers=(1 4 8 12 16 20 24 28 32)
     1.7 +cpus=(0 3 7 15 15 23 23 31 31)
     1.8 +nodes=(0 0 0 1 1 2 2 3 3)
     1.9 +
    1.10 +confs=( "1 1 1" "1 2 2" "2 3 4" "2 4 5" "3 5 8" "3 6 10" "4 7 12" "4 8 15" "5 8 17"		#small
    1.11 +		"1 1 1" "1 2 2" "2 3 4" "2 4 5" "3 5 7" "3 6 9" "4 7 12" "4 8 13" "5 10 15")    #large
    1.12 +
    1.13 +
    1.14 +
    1.15 +#confsmall=("1 1 1" "1 2 2" "2 3 4" "2 4 5" "3 5 8" "3 6 10" "4 7 12" "4 8 15" "5 8 17")
    1.16 +# "7 10 21" "8 12 25" "10 15 29" "11 17 32")
    1.17 +#conflarge=("1 1 1" "1 2 2" "2 3 4" "2 4 5" "3 5 7" "3 6 9" "4 7 12" "4 8 13" "5 10 15")
    1.18 +#"5 12 21" "6 15 25" "7 17 30" "8 19 36")
    1.19 +
    1.20 +
    1.21 +configs=9
    1.22 +
    1.23 +average_ompss_2d=0
    1.24 +average_ompss_3d=0
    1.25 +average_pthread=0
    1.26 +average_serial=0
    1.27 +
    1.28 +iterations_low=4
    1.29 +iterations_high=8
    1.30 +
    1.31 +nframes=10000  # max frames limit for debug purpose
    1.32 +inputs=("14" "10")
    1.33 +inputs_vebose=("Big Bug Bunny 1920x1080 10000 frames" "Park Joy 3840x2160 2500 frames")
    1.34 +osargs=("-z 8 8" "-z 12 12 --static-3d")
    1.35 +
    1.36 +time_stamp=`date +%Y.%m.%d_%H.%M.%S`
    1.37 +outputdir="/home/stefan.hauser/ffmpeg_smp/ppopp_results/rx600s5-1t/$time_stamp"
    1.38 +ompss_2d="$outputdir/ompss_2d.txt"
    1.39 +ompss_3d="$outputdir/ompss_3d.txt"
    1.40 +pthread="$outputdir/pthread.txt"
    1.41 +serial="$outputdir/serial.txt"
    1.42 +
    1.43 +#executes the experiments for a single conf $1=confnum $2 iterations $3 input_idx
    1.44 +function execute_single_conf {
    1.45 +	conf=$1
    1.46 +	iter=$2
    1.47 +	iidx=$3
    1.48 +
    1.49 +	average_ompss_2d=0
    1.50 +	average_ompss_3d=0
    1.51 +	average_pthread=0
    1.52 +
    1.53 +	echo "Workers: " ${workers[$conf]} | tee -a $ompss_2d $ompss_3d $pthread $serial
    1.54 +
    1.55 +	cd build-ss
    1.56 +	for ((i=1;i<=$iter;i+=1)); do
    1.57 +	    # OMPSS
    1.58 +	    #export CSS_NUM_CPUS=$worker
    1.59 +	    NX_PES=${workers[$conf]} numactl --interleave=0-${nodes[$conf]} time -p ./ffmpeg -i ${inputs[$iidx]} -n $nframes -e $((${workers[$conf]}+1)) ${osargs[0]} 2> output
    1.60 +		runtime=$(cat output | grep real | sed s/^.*l.//g)
    1.61 +	    average_ompss_2d=$(echo "$average_ompss_2d + $runtime"|bc)
    1.62 +	    echo -n $runtime " " >> $ompss_2d
    1.63 +	done
    1.64 +
    1.65 +	for ((i=1;i<=$iter;i+=1)); do
    1.66 +		NX_PES=${workers[$conf]} numactl --interleave=0-${nodes[$conf]} time -p ./ffmpeg -i ${inputs[$iidx]} -n $nframes -e $((${workers[$conf]}+1)) ${osargs[1]} 2> output
    1.67 +		runtime=$(cat output | grep real | sed s/^.*l.//g)
    1.68 +	    average_ompss_3d=$(echo "$average_ompss_3d + $runtime"|bc)
    1.69 +	    echo -n $runtime " " >> $ompss_3d
    1.70 +	done
    1.71 +	cd ..
    1.72 +
    1.73 +	cd build
    1.74 +	for ((i=1;i<=$iter;i+=1)); do
    1.75 +		# Pthreads
    1.76 +	    numactl --physcpubind=0-$((${cpus[$conf]})) time -p ./ffmpeg -i ${inputs[$iidx]} -n $nframes -t ${confs[$(($conf + $iidx * $configs))]} 2> output
    1.77 +		runtime=$(cat output | grep real | sed s/^.*l.//g)
    1.78 +	    average_pthread=$(echo "$average_pthread + $runtime"|bc)
    1.79 +	    echo -n $runtime " " >> $pthread
    1.80 +	done
    1.81 +	cd ..
    1.82 +
    1.83 +	echo "" | tee -a $pthread $ompss_2d $ompss_3d
    1.84 +	average_ompss_2d=$(echo "scale=5;$average_ompss_2d/$iter"|bc)
    1.85 +	average_ompss_3d=$(echo "scale=5;$average_ompss_3d/$iter"|bc)
    1.86 +	average_pthread=$(echo "scale=5;$average_pthread/$iter"|bc)
    1.87 +
    1.88 +	echo "time: " $average_ompss_2d >> $ompss_2d
    1.89 +	echo "time: " $average_ompss_3d >> $ompss_3d
    1.90 +	echo "time: " $average_pthread >> $pthread
    1.91 +	echo "time: " $average_serial >> $serial
    1.92 +}
    1.93 +
    1.94 +
    1.95 +mkdir $outputdir
    1.96 +
    1.97 +echo "Processing inputs ..."
    1.98 +
    1.99 +echo "h264dec Benchmark" | tee $ompss_2d $ompss_3d $pthread $serial
   1.100 +
   1.101 +for n in 0 1; do
   1.102 +	echo "Input: ${inputs_vebose[$n]}" | tee -a $ompss_2d $ompss_3d $pthread $serial
   1.103 +	echo "" | tee -a $ompss_2d $ompss_3d $pthread $serial
   1.104 +
   1.105 +	# Serial
   1.106 +	cd build
   1.107 +	numactl --physcpubind=0 time -p ./ffmpeg -i ${inputs[$n]} -n $nframes -s 2> output
   1.108 +	runtime=$(cat output | grep real | sed s/^.*l.//g)
   1.109 +	average_serial=$runtime
   1.110 +	cd ..
   1.111 +
   1.112 +	execute_single_conf 0 1 $n
   1.113 +
   1.114 +	#Parallel
   1.115 +	for ((confidx=1;confidx<=4;confidx+=1)); do
   1.116 +		execute_single_conf $confidx $iterations_low $n		
   1.117 +	done
   1.118 +
   1.119 +	for ((confidx=5;confidx<=$(($configs-1));confidx+=1)); do
   1.120 +		execute_single_conf $confidx $iterations_high $n		
   1.121 +	done
   1.122 +
   1.123 +	echo "-------------------" | tee -a $ompss_2d $ompss_3d $pthread $serial
   1.124 +done
   1.125 +
   1.126 +echo "FINISHED"
   1.127 +
   1.128 +rm build/output build-ss/output
   1.129 +

     2.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     2.2 +++ b/ffmpeg_smp/h264dec/COPYING.GPLv3	Mon Aug 27 12:09:56 2012 +0200
     2.3 @@ -0,0 +1,674 @@
     2.4 +                    GNU GENERAL PUBLIC LICENSE
     2.5 +                       Version 3, 29 June 2007
     2.6 +
     2.7 + Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/>
     2.8 + Everyone is permitted to copy and distribute verbatim copies
     2.9 + of this license document, but changing it is not allowed.
    2.10 +
    2.11 +                            Preamble
    2.12 +
    2.13 +  The GNU General Public License is a free, copyleft license for
    2.14 +software and other kinds of works.
    2.15 +
    2.16 +  The licenses for most software and other practical works are designed
    2.17 +to take away your freedom to share and change the works.  By contrast,
    2.18 +the GNU General Public License is intended to guarantee your freedom to
    2.19 +share and change all versions of a program--to make sure it remains free
    2.20 +software for all its users.  We, the Free Software Foundation, use the
    2.21 +GNU General Public License for most of our software; it applies also to
    2.22 +any other work released this way by its authors.  You can apply it to
    2.23 +your programs, too.
    2.24 +
    2.25 +  When we speak of free software, we are referring to freedom, not
    2.26 +price.  Our General Public Licenses are designed to make sure that you
    2.27 +have the freedom to distribute copies of free software (and charge for
    2.28 +them if you wish), that you receive source code or can get it if you
    2.29 +want it, that you can change the software or use pieces of it in new
    2.30 +free programs, and that you know you can do these things.
    2.31 +
    2.32 +  To protect your rights, we need to prevent others from denying you
    2.33 +these rights or asking you to surrender the rights.  Therefore, you have
    2.34 +certain responsibilities if you distribute copies of the software, or if
    2.35 +you modify it: responsibilities to respect the freedom of others.
    2.36 +
    2.37 +  For example, if you distribute copies of such a program, whether
    2.38 +gratis or for a fee, you must pass on to the recipients the same
    2.39 +freedoms that you received.  You must make sure that they, too, receive
    2.40 +or can get the source code.  And you must show them these terms so they
    2.41 +know their rights.
    2.42 +
    2.43 +  Developers that use the GNU GPL protect your rights with two steps:
    2.44 +(1) assert copyright on the software, and (2) offer you this License
    2.45 +giving you legal permission to copy, distribute and/or modify it.
    2.46 +
    2.47 +  For the developers' and authors' protection, the GPL clearly explains
    2.48 +that there is no warranty for this free software.  For both users' and
    2.49 +authors' sake, the GPL requires that modified versions be marked as
    2.50 +changed, so that their problems will not be attributed erroneously to
    2.51 +authors of previous versions.
    2.52 +
    2.53 +  Some devices are designed to deny users access to install or run
    2.54 +modified versions of the software inside them, although the manufacturer
    2.55 +can do so.  This is fundamentally incompatible with the aim of
    2.56 +protecting users' freedom to change the software.  The systematic
    2.57 +pattern of such abuse occurs in the area of products for individuals to
    2.58 +use, which is precisely where it is most unacceptable.  Therefore, we
    2.59 +have designed this version of the GPL to prohibit the practice for those
    2.60 +products.  If such problems arise substantially in other domains, we
    2.61 +stand ready to extend this provision to those domains in future versions
    2.62 +of the GPL, as needed to protect the freedom of users.
    2.63 +
    2.64 +  Finally, every program is threatened constantly by software patents.
    2.65 +States should not allow patents to restrict development and use of
    2.66 +software on general-purpose computers, but in those that do, we wish to
    2.67 +avoid the special danger that patents applied to a free program could
    2.68 +make it effectively proprietary.  To prevent this, the GPL assures that
    2.69 +patents cannot be used to render the program non-free.
    2.70 +
    2.71 +  The precise terms and conditions for copying, distribution and
    2.72 +modification follow.
    2.73 +
    2.74 +                       TERMS AND CONDITIONS
    2.75 +
    2.76 +  0. Definitions.
    2.77 +
    2.78 +  "This License" refers to version 3 of the GNU General Public License.
    2.79 +
    2.80 +  "Copyright" also means copyright-like laws that apply to other kinds of
    2.81 +works, such as semiconductor masks.
    2.82 +
    2.83 +  "The Program" refers to any copyrightable work licensed under this
    2.84 +License.  Each licensee is addressed as "you".  "Licensees" and
    2.85 +"recipients" may be individuals or organizations.
    2.86 +
    2.87 +  To "modify" a work means to copy from or adapt all or part of the work
    2.88 +in a fashion requiring copyright permission, other than the making of an
    2.89 +exact copy.  The resulting work is called a "modified version" of the
    2.90 +earlier work or a work "based on" the earlier work.
    2.91 +
    2.92 +  A "covered work" means either the unmodified Program or a work based
    2.93 +on the Program.
    2.94 +
    2.95 +  To "propagate" a work means to do anything with it that, without
    2.96 +permission, would make you directly or secondarily liable for
    2.97 +infringement under applicable copyright law, except executing it on a
    2.98 +computer or modifying a private copy.  Propagation includes copying,
    2.99 +distribution (with or without modification), making available to the
   2.100 +public, and in some countries other activities as well.
   2.101 +
   2.102 +  To "convey" a work means any kind of propagation that enables other
   2.103 +parties to make or receive copies.  Mere interaction with a user through
   2.104 +a computer network, with no transfer of a copy, is not conveying.
   2.105 +
   2.106 +  An interactive user interface displays "Appropriate Legal Notices"
   2.107 +to the extent that it includes a convenient and prominently visible
   2.108 +feature that (1) displays an appropriate copyright notice, and (2)
   2.109 +tells the user that there is no warranty for the work (except to the
   2.110 +extent that warranties are provided), that licensees may convey the
   2.111 +work under this License, and how to view a copy of this License.  If
   2.112 +the interface presents a list of user commands or options, such as a
   2.113 +menu, a prominent item in the list meets this criterion.
   2.114 +
   2.115 +  1. Source Code.
   2.116 +
   2.117 +  The "source code" for a work means the preferred form of the work
   2.118 +for making modifications to it.  "Object code" means any non-source
   2.119 +form of a work.
   2.120 +
   2.121 +  A "Standard Interface" means an interface that either is an official
   2.122 +standard defined by a recognized standards body, or, in the case of
   2.123 +interfaces specified for a particular programming language, one that
   2.124 +is widely used among developers working in that language.
   2.125 +
   2.126 +  The "System Libraries" of an executable work include anything, other
   2.127 +than the work as a whole, that (a) is included in the normal form of
   2.128 +packaging a Major Component, but which is not part of that Major
   2.129 +Component, and (b) serves only to enable use of the work with that
   2.130 +Major Component, or to implement a Standard Interface for which an
   2.131 +implementation is available to the public in source code form.  A
   2.132 +"Major Component", in this context, means a major essential component
   2.133 +(kernel, window system, and so on) of the specific operating system
   2.134 +(if any) on which the executable work runs, or a compiler used to
   2.135 +produce the work, or an object code interpreter used to run it.
   2.136 +
   2.137 +  The "Corresponding Source" for a work in object code form means all
   2.138 +the source code needed to generate, install, and (for an executable
   2.139 +work) run the object code and to modify the work, including scripts to
   2.140 +control those activities.  However, it does not include the work's
   2.141 +System Libraries, or general-purpose tools or generally available free
   2.142 +programs which are used unmodified in performing those activities but
   2.143 +which are not part of the work.  For example, Corresponding Source
   2.144 +includes interface definition files associated with source files for
   2.145 +the work, and the source code for shared libraries and dynamically
   2.146 +linked subprograms that the work is specifically designed to require,
   2.147 +such as by intimate data communication or control flow between those
   2.148 +subprograms and other parts of the work.
   2.149 +
   2.150 +  The Corresponding Source need not include anything that users
   2.151 +can regenerate automatically from other parts of the Corresponding
   2.152 +Source.
   2.153 +
   2.154 +  The Corresponding Source for a work in source code form is that
   2.155 +same work.
   2.156 +
   2.157 +  2. Basic Permissions.
   2.158 +
   2.159 +  All rights granted under this License are granted for the term of
   2.160 +copyright on the Program, and are irrevocable provided the stated
   2.161 +conditions are met.  This License explicitly affirms your unlimited
   2.162 +permission to run the unmodified Program.  The output from running a
   2.163 +covered work is covered by this License only if the output, given its
   2.164 +content, constitutes a covered work.  This License acknowledges your
   2.165 +rights of fair use or other equivalent, as provided by copyright law.
   2.166 +
   2.167 +  You may make, run and propagate covered works that you do not
   2.168 +convey, without conditions so long as your license otherwise remains
   2.169 +in force.  You may convey covered works to others for the sole purpose
   2.170 +of having them make modifications exclusively for you, or provide you
   2.171 +with facilities for running those works, provided that you comply with
   2.172 +the terms of this License in conveying all material for which you do
   2.173 +not control copyright.  Those thus making or running the covered works
   2.174 +for you must do so exclusively on your behalf, under your direction
   2.175 +and control, on terms that prohibit them from making any copies of
   2.176 +your copyrighted material outside their relationship with you.
   2.177 +
   2.178 +  Conveying under any other circumstances is permitted solely under
   2.179 +the conditions stated below.  Sublicensing is not allowed; section 10
   2.180 +makes it unnecessary.
   2.181 +
   2.182 +  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
   2.183 +
   2.184 +  No covered work shall be deemed part of an effective technological
   2.185 +measure under any applicable law fulfilling obligations under article
   2.186 +11 of the WIPO copyright treaty adopted on 20 December 1996, or
   2.187 +similar laws prohibiting or restricting circumvention of such
   2.188 +measures.
   2.189 +
   2.190 +  When you convey a covered work, you waive any legal power to forbid
   2.191 +circumvention of technological measures to the extent such circumvention
   2.192 +is effected by exercising rights under this License with respect to
   2.193 +the covered work, and you disclaim any intention to limit operation or
   2.194 +modification of the work as a means of enforcing, against the work's
   2.195 +users, your or third parties' legal rights to forbid circumvention of
   2.196 +technological measures.
   2.197 +
   2.198 +  4. Conveying Verbatim Copies.
   2.199 +
   2.200 +  You may convey verbatim copies of the Program's source code as you
   2.201 +receive it, in any medium, provided that you conspicuously and
   2.202 +appropriately publish on each copy an appropriate copyright notice;
   2.203 +keep intact all notices stating that this License and any
   2.204 +non-permissive terms added in accord with section 7 apply to the code;
   2.205 +keep intact all notices of the absence of any warranty; and give all
   2.206 +recipients a copy of this License along with the Program.
   2.207 +
   2.208 +  You may charge any price or no price for each copy that you convey,
   2.209 +and you may offer support or warranty protection for a fee.
   2.210 +
   2.211 +  5. Conveying Modified Source Versions.
   2.212 +
   2.213 +  You may convey a work based on the Program, or the modifications to
   2.214 +produce it from the Program, in the form of source code under the
   2.215 +terms of section 4, provided that you also meet all of these conditions:
   2.216 +
   2.217 +    a) The work must carry prominent notices stating that you modified
   2.218 +    it, and giving a relevant date.
   2.219 +
   2.220 +    b) The work must carry prominent notices stating that it is
   2.221 +    released under this License and any conditions added under section
   2.222 +    7.  This requirement modifies the requirement in section 4 to
   2.223 +    "keep intact all notices".
   2.224 +
   2.225 +    c) You must license the entire work, as a whole, under this
   2.226 +    License to anyone who comes into possession of a copy.  This
   2.227 +    License will therefore apply, along with any applicable section 7
   2.228 +    additional terms, to the whole of the work, and all its parts,
   2.229 +    regardless of how they are packaged.  This License gives no
   2.230 +    permission to license the work in any other way, but it does not
   2.231 +    invalidate such permission if you have separately received it.
   2.232 +
   2.233 +    d) If the work has interactive user interfaces, each must display
   2.234 +    Appropriate Legal Notices; however, if the Program has interactive
   2.235 +    interfaces that do not display Appropriate Legal Notices, your
   2.236 +    work need not make them do so.
   2.237 +
   2.238 +  A compilation of a covered work with other separate and independent
   2.239 +works, which are not by their nature extensions of the covered work,
   2.240 +and which are not combined with it such as to form a larger program,
   2.241 +in or on a volume of a storage or distribution medium, is called an
   2.242 +"aggregate" if the compilation and its resulting copyright are not
   2.243 +used to limit the access or legal rights of the compilation's users
   2.244 +beyond what the individual works permit.  Inclusion of a covered work
   2.245 +in an aggregate does not cause this License to apply to the other
   2.246 +parts of the aggregate.
   2.247 +
   2.248 +  6. Conveying Non-Source Forms.
   2.249 +
   2.250 +  You may convey a covered work in object code form under the terms
   2.251 +of sections 4 and 5, provided that you also convey the
   2.252 +machine-readable Corresponding Source under the terms of this License,
   2.253 +in one of these ways:
   2.254 +
   2.255 +    a) Convey the object code in, or embodied in, a physical product
   2.256 +    (including a physical distribution medium), accompanied by the
   2.257 +    Corresponding Source fixed on a durable physical medium
   2.258 +    customarily used for software interchange.
   2.259 +
   2.260 +    b) Convey the object code in, or embodied in, a physical product
   2.261 +    (including a physical distribution medium), accompanied by a
   2.262 +    written offer, valid for at least three years and valid for as
   2.263 +    long as you offer spare parts or customer support for that product
   2.264 +    model, to give anyone who possesses the object code either (1) a
   2.265 +    copy of the Corresponding Source for all the software in the
   2.266 +    product that is covered by this License, on a durable physical
   2.267 +    medium customarily used for software interchange, for a price no
   2.268 +    more than your reasonable cost of physically performing this
   2.269 +    conveying of source, or (2) access to copy the
   2.270 +    Corresponding Source from a network server at no charge.
   2.271 +
   2.272 +    c) Convey individual copies of the object code with a copy of the
   2.273 +    written offer to provide the Corresponding Source.  This
   2.274 +    alternative is allowed only occasionally and noncommercially, and
   2.275 +    only if you received the object code with such an offer, in accord
   2.276 +    with subsection 6b.
   2.277 +
   2.278 +    d) Convey the object code by offering access from a designated
   2.279 +    place (gratis or for a charge), and offer equivalent access to the
   2.280 +    Corresponding Source in the same way through the same place at no
   2.281 +    further charge.  You need not require recipients to copy the
   2.282 +    Corresponding Source along with the object code.  If the place to
   2.283 +    copy the object code is a network server, the Corresponding Source
   2.284 +    may be on a different server (operated by you or a third party)
   2.285 +    that supports equivalent copying facilities, provided you maintain
   2.286 +    clear directions next to the object code saying where to find the
   2.287 +    Corresponding Source.  Regardless of what server hosts the
   2.288 +    Corresponding Source, you remain obligated to ensure that it is
   2.289 +    available for as long as needed to satisfy these requirements.
   2.290 +
   2.291 +    e) Convey the object code using peer-to-peer transmission, provided
   2.292 +    you inform other peers where the object code and Corresponding
   2.293 +    Source of the work are being offered to the general public at no
   2.294 +    charge under subsection 6d.
   2.295 +
   2.296 +  A separable portion of the object code, whose source code is excluded
   2.297 +from the Corresponding Source as a System Library, need not be
   2.298 +included in conveying the object code work.
   2.299 +
   2.300 +  A "User Product" is either (1) a "consumer product", which means any
   2.301 +tangible personal property which is normally used for personal, family,
   2.302 +or household purposes, or (2) anything designed or sold for incorporation
   2.303 +into a dwelling.  In determining whether a product is a consumer product,
   2.304 +doubtful cases shall be resolved in favor of coverage.  For a particular
   2.305 +product received by a particular user, "normally used" refers to a
   2.306 +typical or common use of that class of product, regardless of the status
   2.307 +of the particular user or of the way in which the particular user
   2.308 +actually uses, or expects or is expected to use, the product.  A product
   2.309 +is a consumer product regardless of whether the product has substantial
   2.310 +commercial, industrial or non-consumer uses, unless such uses represent
   2.311 +the only significant mode of use of the product.
   2.312 +
   2.313 +  "Installation Information" for a User Product means any methods,
   2.314 +procedures, authorization keys, or other information required to install
   2.315 +and execute modified versions of a covered work in that User Product from
   2.316 +a modified version of its Corresponding Source.  The information must
   2.317 +suffice to ensure that the continued functioning of the modified object
   2.318 +code is in no case prevented or interfered with solely because
   2.319 +modification has been made.
   2.320 +
   2.321 +  If you convey an object code work under this section in, or with, or
   2.322 +specifically for use in, a User Product, and the conveying occurs as
   2.323 +part of a transaction in which the right of possession and use of the
   2.324 +User Product is transferred to the recipient in perpetuity or for a
   2.325 +fixed term (regardless of how the transaction is characterized), the
   2.326 +Corresponding Source conveyed under this section must be accompanied
   2.327 +by the Installation Information.  But this requirement does not apply
   2.328 +if neither you nor any third party retains the ability to install
   2.329 +modified object code on the User Product (for example, the work has
   2.330 +been installed in ROM).
   2.331 +
   2.332 +  The requirement to provide Installation Information does not include a
   2.333 +requirement to continue to provide support service, warranty, or updates
   2.334 +for a work that has been modified or installed by the recipient, or for
   2.335 +the User Product in which it has been modified or installed.  Access to a
   2.336 +network may be denied when the modification itself materially and
   2.337 +adversely affects the operation of the network or violates the rules and
   2.338 +protocols for communication across the network.
   2.339 +
   2.340 +  Corresponding Source conveyed, and Installation Information provided,
   2.341 +in accord with this section must be in a format that is publicly
   2.342 +documented (and with an implementation available to the public in
   2.343 +source code form), and must require no special password or key for
   2.344 +unpacking, reading or copying.
   2.345 +
   2.346 +  7. Additional Terms.
   2.347 +
   2.348 +  "Additional permissions" are terms that supplement the terms of this
   2.349 +License by making exceptions from one or more of its conditions.
   2.350 +Additional permissions that are applicable to the entire Program shall
   2.351 +be treated as though they were included in this License, to the extent
   2.352 +that they are valid under applicable law.  If additional permissions
   2.353 +apply only to part of the Program, that part may be used separately
   2.354 +under those permissions, but the entire Program remains governed by
   2.355 +this License without regard to the additional permissions.
   2.356 +
   2.357 +  When you convey a copy of a covered work, you may at your option
   2.358 +remove any additional permissions from that copy, or from any part of
   2.359 +it.  (Additional permissions may be written to require their own
   2.360 +removal in certain cases when you modify the work.)  You may place
   2.361 +additional permissions on material, added by you to a covered work,
   2.362 +for which you have or can give appropriate copyright permission.
   2.363 +
   2.364 +  Notwithstanding any other provision of this License, for material you
   2.365 +add to a covered work, you may (if authorized by the copyright holders of
   2.366 +that material) supplement the terms of this License with terms:
   2.367 +
   2.368 +    a) Disclaiming warranty or limiting liability differently from the
   2.369 +    terms of sections 15 and 16 of this License; or
   2.370 +
   2.371 +    b) Requiring preservation of specified reasonable legal notices or
   2.372 +    author attributions in that material or in the Appropriate Legal
   2.373 +    Notices displayed by works containing it; or
   2.374 +
   2.375 +    c) Prohibiting misrepresentation of the origin of that material, or
   2.376 +    requiring that modified versions of such material be marked in
   2.377 +    reasonable ways as different from the original version; or
   2.378 +
   2.379 +    d) Limiting the use for publicity purposes of names of licensors or
   2.380 +    authors of the material; or
   2.381 +
   2.382 +    e) Declining to grant rights under trademark law for use of some
   2.383 +    trade names, trademarks, or service marks; or
   2.384 +
   2.385 +    f) Requiring indemnification of licensors and authors of that
   2.386 +    material by anyone who conveys the material (or modified versions of
   2.387 +    it) with contractual assumptions of liability to the recipient, for
   2.388 +    any liability that these contractual assumptions directly impose on
   2.389 +    those licensors and authors.
   2.390 +
   2.391 +  All other non-permissive additional terms are considered "further
   2.392 +restrictions" within the meaning of section 10.  If the Program as you
   2.393 +received it, or any part of it, contains a notice stating that it is
   2.394 +governed by this License along with a term that is a further
   2.395 +restriction, you may remove that term.  If a license document contains
   2.396 +a further restriction but permits relicensing or conveying under this
   2.397 +License, you may add to a covered work material governed by the terms
   2.398 +of that license document, provided that the further restriction does
   2.399 +not survive such relicensing or conveying.
   2.400 +
   2.401 +  If you add terms to a covered work in accord with this section, you
   2.402 +must place, in the relevant source files, a statement of the
   2.403 +additional terms that apply to those files, or a notice indicating
   2.404 +where to find the applicable terms.
   2.405 +
   2.406 +  Additional terms, permissive or non-permissive, may be stated in the
   2.407 +form of a separately written license, or stated as exceptions;
   2.408 +the above requirements apply either way.
   2.409 +
   2.410 +  8. Termination.
   2.411 +
   2.412 +  You may not propagate or modify a covered work except as expressly
   2.413 +provided under this License.  Any attempt otherwise to propagate or
   2.414 +modify it is void, and will automatically terminate your rights under
   2.415 +this License (including any patent licenses granted under the third
   2.416 +paragraph of section 11).
   2.417 +
   2.418 +  However, if you cease all violation of this License, then your
   2.419 +license from a particular copyright holder is reinstated (a)
   2.420 +provisionally, unless and until the copyright holder explicitly and
   2.421 +finally terminates your license, and (b) permanently, if the copyright
   2.422 +holder fails to notify you of the violation by some reasonable means
   2.423 +prior to 60 days after the cessation.
   2.424 +
   2.425 +  Moreover, your license from a particular copyright holder is
   2.426 +reinstated permanently if the copyright holder notifies you of the
   2.427 +violation by some reasonable means, this is the first time you have
   2.428 +received notice of violation of this License (for any work) from that
   2.429 +copyright holder, and you cure the violation prior to 30 days after
   2.430 +your receipt of the notice.
   2.431 +
   2.432 +  Termination of your rights under this section does not terminate the
   2.433 +licenses of parties who have received copies or rights from you under
   2.434 +this License.  If your rights have been terminated and not permanently
   2.435 +reinstated, you do not qualify to receive new licenses for the same
   2.436 +material under section 10.
   2.437 +
   2.438 +  9. Acceptance Not Required for Having Copies.
   2.439 +
   2.440 +  You are not required to accept this License in order to receive or
   2.441 +run a copy of the Program.  Ancillary propagation of a covered work
   2.442 +occurring solely as a consequence of using peer-to-peer transmission
   2.443 +to receive a copy likewise does not require acceptance.  However,
   2.444 +nothing other than this License grants you permission to propagate or
   2.445 +modify any covered work.  These actions infringe copyright if you do
   2.446 +not accept this License.  Therefore, by modifying or propagating a
   2.447 +covered work, you indicate your acceptance of this License to do so.
   2.448 +
   2.449 +  10. Automatic Licensing of Downstream Recipients.
   2.450 +
   2.451 +  Each time you convey a covered work, the recipient automatically
   2.452 +receives a license from the original licensors, to run, modify and
   2.453 +propagate that work, subject to this License.  You are not responsible
   2.454 +for enforcing compliance by third parties with this License.
   2.455 +
   2.456 +  An "entity transaction" is a transaction transferring control of an
   2.457 +organization, or substantially all assets of one, or subdividing an
   2.458 +organization, or merging organizations.  If propagation of a covered
   2.459 +work results from an entity transaction, each party to that
   2.460 +transaction who receives a copy of the work also receives whatever
   2.461 +licenses to the work the party's predecessor in interest had or could
   2.462 +give under the previous paragraph, plus a right to possession of the
   2.463 +Corresponding Source of the work from the predecessor in interest, if
   2.464 +the predecessor has it or can get it with reasonable efforts.
   2.465 +
   2.466 +  You may not impose any further restrictions on the exercise of the
   2.467 +rights granted or affirmed under this License.  For example, you may
   2.468 +not impose a license fee, royalty, or other charge for exercise of
   2.469 +rights granted under this License, and you may not initiate litigation
   2.470 +(including a cross-claim or counterclaim in a lawsuit) alleging that
   2.471 +any patent claim is infringed by making, using, selling, offering for
   2.472 +sale, or importing the Program or any portion of it.
   2.473 +
   2.474 +  11. Patents.
   2.475 +
   2.476 +  A "contributor" is a copyright holder who authorizes use under this
   2.477 +License of the Program or a work on which the Program is based.  The
   2.478 +work thus licensed is called the contributor's "contributor version".
   2.479 +
   2.480 +  A contributor's "essential patent claims" are all patent claims
   2.481 +owned or controlled by the contributor, whether already acquired or
   2.482 +hereafter acquired, that would be infringed by some manner, permitted
   2.483 +by this License, of making, using, or selling its contributor version,
   2.484 +but do not include claims that would be infringed only as a
   2.485 +consequence of further modification of the contributor version.  For
   2.486 +purposes of this definition, "control" includes the right to grant
   2.487 +patent sublicenses in a manner consistent with the requirements of
   2.488 +this License.
   2.489 +
   2.490 +  Each contributor grants you a non-exclusive, worldwide, royalty-free
   2.491 +patent license under the contributor's essential patent claims, to
   2.492 +make, use, sell, offer for sale, import and otherwise run, modify and
   2.493 +propagate the contents of its contributor version.
   2.494 +
   2.495 +  In the following three paragraphs, a "patent license" is any express
   2.496 +agreement or commitment, however denominated, not to enforce a patent
   2.497 +(such as an express permission to practice a patent or covenant not to
   2.498 +sue for patent infringement).  To "grant" such a patent license to a
   2.499 +party means to make such an agreement or commitment not to enforce a
   2.500 +patent against the party.
   2.501 +
   2.502 +  If you convey a covered work, knowingly relying on a patent license,
   2.503 +and the Corresponding Source of the work is not available for anyone
   2.504 +to copy, free of charge and under the terms of this License, through a
   2.505 +publicly available network server or other readily accessible means,
   2.506 +then you must either (1) cause the Corresponding Source to be so
   2.507 +available, or (2) arrange to deprive yourself of the benefit of the
   2.508 +patent license for this particular work, or (3) arrange, in a manner
   2.509 +consistent with the requirements of this License, to extend the patent
   2.510 +license to downstream recipients.  "Knowingly relying" means you have
   2.511 +actual knowledge that, but for the patent license, your conveying the
   2.512 +covered work in a country, or your recipient's use of the covered work
   2.513 +in a country, would infringe one or more identifiable patents in that
   2.514 +country that you have reason to believe are valid.
   2.515 +
   2.516 +  If, pursuant to or in connection with a single transaction or
   2.517 +arrangement, you convey, or propagate by procuring conveyance of, a
   2.518 +covered work, and grant a patent license to some of the parties
   2.519 +receiving the covered work authorizing them to use, propagate, modify
   2.520 +or convey a specific copy of the covered work, then the patent license
   2.521 +you grant is automatically extended to all recipients of the covered
   2.522 +work and works based on it.
   2.523 +
   2.524 +  A patent license is "discriminatory" if it does not include within
   2.525 +the scope of its coverage, prohibits the exercise of, or is
   2.526 +conditioned on the non-exercise of one or more of the rights that are
   2.527 +specifically granted under this License.  You may not convey a covered
   2.528 +work if you are a party to an arrangement with a third party that is
   2.529 +in the business of distributing software, under which you make payment
   2.530 +to the third party based on the extent of your activity of conveying
   2.531 +the work, and under which the third party grants, to any of the
   2.532 +parties who would receive the covered work from you, a discriminatory
   2.533 +patent license (a) in connection with copies of the covered work
   2.534 +conveyed by you (or copies made from those copies), or (b) primarily
   2.535 +for and in connection with specific products or compilations that
   2.536 +contain the covered work, unless you entered into that arrangement,
   2.537 +or that patent license was granted, prior to 28 March 2007.
   2.538 +
   2.539 +  Nothing in this License shall be construed as excluding or limiting
   2.540 +any implied license or other defenses to infringement that may
   2.541 +otherwise be available to you under applicable patent law.
   2.542 +
   2.543 +  12. No Surrender of Others' Freedom.
   2.544 +
   2.545 +  If conditions are imposed on you (whether by court order, agreement or
   2.546 +otherwise) that contradict the conditions of this License, they do not
   2.547 +excuse you from the conditions of this License.  If you cannot convey a
   2.548 +covered work so as to satisfy simultaneously your obligations under this
   2.549 +License and any other pertinent obligations, then as a consequence you may
   2.550 +not convey it at all.  For example, if you agree to terms that obligate you
   2.551 +to collect a royalty for further conveying from those to whom you convey
   2.552 +the Program, the only way you could satisfy both those terms and this
   2.553 +License would be to refrain entirely from conveying the Program.
   2.554 +
   2.555 +  13. Use with the GNU Affero General Public License.
   2.556 +
   2.557 +  Notwithstanding any other provision of this License, you have
   2.558 +permission to link or combine any covered work with a work licensed
   2.559 +under version 3 of the GNU Affero General Public License into a single
   2.560 +combined work, and to convey the resulting work.  The terms of this
   2.561 +License will continue to apply to the part which is the covered work,
   2.562 +but the special requirements of the GNU Affero General Public License,
   2.563 +section 13, concerning interaction through a network will apply to the
   2.564 +combination as such.
   2.565 +
   2.566 +  14. Revised Versions of this License.
   2.567 +
   2.568 +  The Free Software Foundation may publish revised and/or new versions of
   2.569 +the GNU General Public License from time to time.  Such new versions will
   2.570 +be similar in spirit to the present version, but may differ in detail to
   2.571 +address new problems or concerns.
   2.572 +
   2.573 +  Each version is given a distinguishing version number.  If the
   2.574 +Program specifies that a certain numbered version of the GNU General
   2.575 +Public License "or any later version" applies to it, you have the
   2.576 +option of following the terms and conditions either of that numbered
   2.577 +version or of any later version published by the Free Software
   2.578 +Foundation.  If the Program does not specify a version number of the
   2.579 +GNU General Public License, you may choose any version ever published
   2.580 +by the Free Software Foundation.
   2.581 +
   2.582 +  If the Program specifies that a proxy can decide which future
   2.583 +versions of the GNU General Public License can be used, that proxy's
   2.584 +public statement of acceptance of a version permanently authorizes you
   2.585 +to choose that version for the Program.
   2.586 +
   2.587 +  Later license versions may give you additional or different
   2.588 +permissions.  However, no additional obligations are imposed on any
   2.589 +author or copyright holder as a result of your choosing to follow a
   2.590 +later version.
   2.591 +
   2.592 +  15. Disclaimer of Warranty.
   2.593 +
   2.594 +  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
   2.595 +APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
   2.596 +HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
   2.597 +OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
   2.598 +THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
   2.599 +PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
   2.600 +IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
   2.601 +ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
   2.602 +
   2.603 +  16. Limitation of Liability.
   2.604 +
   2.605 +  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
   2.606 +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
   2.607 +THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
   2.608 +GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
   2.609 +USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
   2.610 +DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
   2.611 +PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
   2.612 +EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
   2.613 +SUCH DAMAGES.
   2.614 +
   2.615 +  17. Interpretation of Sections 15 and 16.
   2.616 +
   2.617 +  If the disclaimer of warranty and limitation of liability provided
   2.618 +above cannot be given local legal effect according to their terms,
   2.619 +reviewing courts shall apply local law that most closely approximates
   2.620 +an absolute waiver of all civil liability in connection with the
   2.621 +Program, unless a warranty or assumption of liability accompanies a
   2.622 +copy of the Program in return for a fee.
   2.623 +
   2.624 +                     END OF TERMS AND CONDITIONS
   2.625 +
   2.626 +            How to Apply These Terms to Your New Programs
   2.627 +
   2.628 +  If you develop a new program, and you want it to be of the greatest
   2.629 +possible use to the public, the best way to achieve this is to make it
   2.630 +free software which everyone can redistribute and change under these terms.
   2.631 +
   2.632 +  To do so, attach the following notices to the program.  It is safest
   2.633 +to attach them to the start of each source file to most effectively
   2.634 +state the exclusion of warranty; and each file should have at least
   2.635 +the "copyright" line and a pointer to where the full notice is found.
   2.636 +
   2.637 +    <one line to give the program's name and a brief idea of what it does.>
   2.638 +    Copyright (C) <year>  <name of author>
   2.639 +
   2.640 +    This program is free software: you can redistribute it and/or modify
   2.641 +    it under the terms of the GNU General Public License as published by
   2.642 +    the Free Software Foundation, either version 3 of the License, or
   2.643 +    (at your option) any later version.
   2.644 +
   2.645 +    This program is distributed in the hope that it will be useful,
   2.646 +    but WITHOUT ANY WARRANTY; without even the implied warranty of
   2.647 +    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   2.648 +    GNU General Public License for more details.
   2.649 +
   2.650 +    You should have received a copy of the GNU General Public License
   2.651 +    along with this program.  If not, see <http://www.gnu.org/licenses/>.
   2.652 +
   2.653 +Also add information on how to contact you by electronic and paper mail.
   2.654 +
   2.655 +  If the program does terminal interaction, make it output a short
   2.656 +notice like this when it starts in an interactive mode:
   2.657 +
   2.658 +    <program>  Copyright (C) <year>  <name of author>
   2.659 +    This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
   2.660 +    This is free software, and you are welcome to redistribute it
   2.661 +    under certain conditions; type `show c' for details.
   2.662 +
   2.663 +The hypothetical commands `show w' and `show c' should show the appropriate
   2.664 +parts of the General Public License.  Of course, your program's commands
   2.665 +might be different; for a GUI interface, you would use an "about box".
   2.666 +
   2.667 +  You should also get your employer (if you work as a programmer) or school,
   2.668 +if any, to sign a "copyright disclaimer" for the program, if necessary.
   2.669 +For more information on this, and how to apply and follow the GNU GPL, see
   2.670 +<http://www.gnu.org/licenses/>.
   2.671 +
   2.672 +  The GNU General Public License does not permit incorporating your program
   2.673 +into proprietary programs.  If your program is a subroutine library, you
   2.674 +may consider it more useful to permit linking proprietary applications with
   2.675 +the library.  If this is what you want to do, use the GNU Lesser General
   2.676 +Public License instead of this License.  But first, please read
   2.677 +<http://www.gnu.org/philosophy/why-not-lgpl.html>.

     3.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     3.2 +++ b/ffmpeg_smp/h264dec/README.txt	Mon Aug 27 12:09:56 2012 +0200
     3.3 @@ -0,0 +1,79 @@
     3.4 +App: h264dec
     3.5 +
     3.6 +This application decodes H.264 raw videos.
     3.7 +
     3.8 +Build Sequential/Pthreads:
     3.9 +
    3.10 +autoreconf -i -f
    3.11 +mkdir build
    3.12 +cd build
    3.13 +../configure --enable-ssse3 --enable-sdl2
    3.14 +make
    3.15 +
    3.16 +Build OmpSs:
    3.17 +
    3.18 +autoreconf -i -f
    3.19 +mkdir build
    3.20 +cd build-ss
    3.21 +../configure CC=sscc --enable-ssse3 --enable-sdl2
    3.22 +make
    3.23 +
    3.24 +ssse3 enables assembler optimizations up to ssse3 (optional)
    3.25 +sdl enables a rudimentary viewing capability      (optional)
    3.26 +
    3.27 +Usage Sequential/Pthreads:
    3.28 +./h264dec -i $(INPUT_VIDEO) -s
    3.29 +./h264dec -i $(INPUT_VIDEO) -t $(THREADS)
    3.30 +
    3.31 +Usage OmpSs:
    3.32 +NX_PES=<numthreads> ./h264dec -i <inputfile> -e <num parallel entropy frames> -z <width> <height> --static-3d
    3.33 +
    3.34 +-e specify the number of entropy decode pipeline buffers and should be ideally
    3.35 +the same as the number of threads.
    3.36 +
    3.37 +-z allows to set the MB reconstruction grouped block size. A size between 6 by 6 to 10 by 10
    3.38 +was found to strike a good balance between overhead and parallelism, but is machine and input
    3.39 +dependent.
    3.40 +
    3.41 +--static-3d performs overlapping wavefront decoding.
    3.42 +
    3.43 +General usage:
    3.44 +-d 				displays output
    3.45 +-f 				fullscreen
    3.46 +-o $(OUT_FILE)  write raw YUV
    3.47 +-v  			show framerate
    3.48 +
    3.49 +
    3.50 +The INPUT_VIDEOs are in "inputs_encore", but should be able to decode any raw H.264 stream using
    3.51 +one slice per frame, non-interlaced, and CABAC, YUV420.
    3.52 +
    3.53 +
    3.54 +Integrated OmpSs player demo
    3.55 +----------------------------
    3.56 +NOTE: for the player demo SDL2 must be installed.
    3.57 +
    3.58 +1. Go to the OmpSs build directory (/home/cchi/Projects/ffmpeg_smp/build-ss)
    3.59 +
    3.60 +2. Launch the H.264 decoder with the desired options:
    3.61 +
    3.62 +NX_PES=<numthreads> ./h264dec <inputfile> -v (verbose) -e <num parallel entropy frames> -z <width> <height> -d (display) -f (fullscreen)
    3.63 +
    3.64 +note that <num parallel entropy frames> should be equal or higher than <numthreads> for optimal performance
    3.65 +
    3.66 +Examples:
    3.67 +
    3.68 +NX_PES=7 ./h264dec -i ../../h264_movies/park_joy_2160px5.h264 -v -z 8 8 -df -e 9
    3.69 +NX_PES=7 ./h264dec -i ../../h264_movies/big_buck_bunny_1080p24.h264 -v -d  -z 6 6 -e 9
    3.70 +
    3.71 +Interacting with the program
    3.72 +----------------------------
    3.73 +<CTRL+F>    Fullscreen mode
    3.74 +<ESCAPE>    Window mode
    3.75 +<SPACE>     Pause/resume
    3.76 +<M>         Show/hide macroblock borders
    3.77 +<arrows>    When macroblock borders are shown resizes the macroblocks
    3.78 +<ALT+F4>    Close
    3.79 +
    3.80 +Force close in case of lockup
    3.81 +-----------------------------
    3.82 +On a terminal: killall -9 h264dec

     4.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     4.2 +++ b/ffmpeg_smp/h264dec/configure.ac	Mon Aug 27 12:09:56 2012 +0200
     4.3 @@ -0,0 +1,171 @@
     4.4 +#                                               -*- Autoconf -*-
     4.5 +# Process this file with autoconf to produce a configure script.
     4.6 +
     4.7 +AC_PREREQ(2.61)
     4.8 +AC_INIT([h264_mt], [0.1], [cchi@cs.tu-berlin.de])
     4.9 +#AM_INIT_AUTOMAKE(AC_PACKAGE_NAME, AC_PACKAGE_VERSION)
    4.10 +AM_INIT_AUTOMAKE([-Wall -Werror foreign])
    4.11 +
    4.12 +AC_CONFIG_SRCDIR([h264dec.c])
    4.13 +AC_PROG_RANLIB
    4.14 +
    4.15 +# Checks for programs.
    4.16 +AC_GNU_SOURCE
    4.17 +AC_PROG_CC
    4.18 +AM_CONDITIONAL([HAVE_OMPSS], [test $CC = "sscc"])
    4.19 +AC_DEFINE([OMPSS], [0], [Define to 1 on when using the OmpSs compiler sscc])
    4.20 +if test $CC = "sscc";then
    4.21 +AC_DEFINE([OMPSS], [1], [Define to 1 on when using the OmpSs compiler sscc])
    4.22 +fi
    4.23 +
    4.24 +#if [ test -n "${CFLAGS+x}" ] ; then
    4.25 +#    CFLAGS="-O3 -g"
    4.26 +#fi
    4.27 +
    4.28 +# Checks for libraries.
    4.29 +AC_CHECK_LIB([pthread], [pthread_yield])
    4.30 +AC_CHECK_LIB([spe2], [spe_image_open])
    4.31 +AC_CHECK_LIB([sync], [mutex_init])
    4.32 +AC_CHECK_LIB([rt], [clock_gettime])
    4.33 +
    4.34 +AC_ARG_ENABLE([sdl2], AS_HELP_STRING([--enable-sdl2], [Enable SDL2 playback]))
    4.35 +if test "$enable_sdl2" = "yes"; then
    4.36 +	AC_CHECK_LIB([SDL2], [SDL_CreateWindow], [], [echo "Error! libSDL2 required for playback." exit -1])
    4.37 +fi
    4.38 +
    4.39 +if test "$enable_sdl2" = "yes"; then
    4.40 +	AC_CHECK_LIB([X11], [XInitThreads], [], [echo "Error! libX11 currently required for SDL2 workaround." exit -1])
    4.41 +fi
    4.42 +
    4.43 +AC_ARG_ENABLE([sdl_ttf], AS_HELP_STRING([--enable-sdl_ttf], [Enable SDL_ttf for overlaying fonts]))
    4.44 +if test "$enable_sdl_ttf" = "yes"; then
    4.45 +    AC_CHECK_LIB([SDL_ttf], [TTF_Init], [], [echo "Error! libSDL_ttf required for font rendering." exit -1])
    4.46 +fi
    4.47 +
    4.48 +
    4.49 +
    4.50 +AC_ARG_ENABLE([opencl], AS_HELP_STRING([--enable-opencl], [Enable GPU decoder]))
    4.51 +if test "$enable_opencl" = "yes"; then
    4.52 +	AC_CHECK_LIB([OpenCL], [clGetPlatformIDs], [], [echo "Error! libOpenCL required for GPU functionality." exit -1])
    4.53 +fi
    4.54 +AM_CONDITIONAL([HAVE_OPENCL], [test "$enable_opencl" = "yes"])
    4.55 +
    4.56 +
    4.57 +# Checks for header files.
    4.58 +AC_HEADER_STDC
    4.59 +AC_CHECK_HEADERS([stdint.h stdlib.h string.h unistd.h])
    4.60 +
    4.61 +# Checks for typedefs, structures, and compiler characteristics.
    4.62 +AC_C_CONST
    4.63 +AC_TYPE_UINT32_T
    4.64 +AC_TYPE_UINT64_T
    4.65 +AC_TYPE_UINT8_T
    4.66 +AC_C_VOLATILE
    4.67 +AC_C_BIGENDIAN
    4.68 +
    4.69 +# Checks for library functions.
    4.70 +AC_CHECK_FUNCS([malloc realloc memalign posix_memalign memmove memset])
    4.71 +
    4.72 +AC_CANONICAL_HOST
    4.73 +AC_CANONICAL_BUILD
    4.74 +
    4.75 +AC_MSG_CHECKING([for architecture])
    4.76 +
    4.77 +AC_DEFINE([ARCH_ARM], [0], [Define to 1 on arm architectures.])
    4.78 +AC_DEFINE([ARCH_X86_32], [0], [Define to 1 on x86 architectures.])
    4.79 +AC_DEFINE([ARCH_X86_64], [0], [Define to 1 on x86_64 architectures.])
    4.80 +AC_DEFINE([ARCH_X86], [ARCH_X86_32 ||ARCH_X86_64], [True on x86])
    4.81 +AC_DEFINE([ARCH_PPC], [0], [Define to 1 on ppc architectures.])
    4.82 +AC_DEFINE([ARCH_PPC64], [0], [Define to 1 on ppc64 architectures.])
    4.83 +AC_DEFINE([ARCH_CELL], [0], [Define to 1 on cell architectures.])
    4.84 +
    4.85 +if test "$enable_optimizations" != "no"; then
    4.86 +	case $build_cpu in
    4.87 +		arm )
    4.88 +			arch="arm"
    4.89 +			AC_MSG_RESULT([arm])
    4.90 +			AC_DEFINE([ARCH_ARM], [1], [Define to 1 on arm architectures.])
    4.91 +			;;
    4.92 +		i686 )
    4.93 +			arch="x86"
    4.94 +			AC_MSG_RESULT([x86])
    4.95 +			AC_DEFINE([ARCH_X86_32], [1], [Define to 1 on x86 architectures.])
    4.96 +			;;
    4.97 +		x86_64 )
    4.98 +			arch="x86_64"
    4.99 +			AC_MSG_RESULT([x86_64])
   4.100 +			AC_DEFINE([ARCH_X86_64], [1], [Define to 1 on x86 architectures.])
   4.101 +			;;
   4.102 +		powerpc64 )
   4.103 +			AC_DEFINE([HAVE_BIGENDIAN], [1], [Define to 1 on bigendian architectures.])
   4.104 +			if grep -E ^cpu /proc/cpuinfo | grep -q Cell ; then
   4.105 +				arch="cell"
   4.106 +				AC_MSG_RESULT([cell])
   4.107 +				AC_DEFINE([ARCH_CELL], [1], [Define to 1 on cell architectures.])
   4.108 +			else
   4.109 +				arch="powerpc64"
   4.110 +				AC_MSG_RESULT([ppc64])
   4.111 +				AC_DEFINE([ARCH_PPC64], [1], [Define to 1 on ppc64 architectures.])
   4.112 +			fi
   4.113 +			;;
   4.114 +		* )
   4.115 +			AC_MSG_RESULT([default (little endian).])
   4.116 +			;;
   4.117 +	esac
   4.118 +fi
   4.119 +
   4.120 +AM_CONDITIONAL([HAVE_CELL], [test $arch = "cell"])
   4.121 +
   4.122 +# Additional options
   4.123 +AC_ARG_ENABLE([optimizations], AS_HELP_STRING([--disable-optimizations], [Disable all architecture specific optimizations. Compiler optimizations are not disabled.]))
   4.124 +
   4.125 +AC_DEFINE([HAVE_SSE], [0], [Define to 1 to enable sse optimizations.])
   4.126 +AC_DEFINE([HAVE_MMX], [0], [Define to 1 to enable mmx optimizations.])
   4.127 +AC_DEFINE([HAVE_MMX2], [0], [Define to 1 to enable mmx2 optimizations.])
   4.128 +AC_DEFINE([HAVE_SSSE3], [0], [Define to 1 to enable ssse3 optimizations.])
   4.129 +AC_DEFINE([HAVE_ALTIVEC], [0], [Define to 1 to enable altivec optimizations.])
   4.130 +AC_DEFINE([HAVE_NEON], [0], [Define to 1 to enable neon optimizations.])
   4.131 +
   4.132 +AC_ARG_ENABLE([ssse3], AS_HELP_STRING([--enable-ssse3], [Enable ssse3 optimizations]))
   4.133 +if test "$enable_ssse3" = "yes"; then
   4.134 +	AC_DEFINE([HAVE_SSSE3], [1], [Define to 1 to enable ssse3 optimizations.])
   4.135 +	AC_DEFINE([HAVE_SSE], [1], [Define to 1 to enable sse optimizations.])
   4.136 +	AC_DEFINE([HAVE_MMX], [1], [Define to 1 to enable mmx optimizations.])
   4.137 +	AC_DEFINE([HAVE_MMX2], [1], [Define to 1 to enable mmx2 optimizations.])
   4.138 +	ARCH_SUBDIR=x86
   4.139 +fi
   4.140 +
   4.141 +AC_ARG_ENABLE([sse], AS_HELP_STRING([--enable-sse], [Enable sse optimizations]))
   4.142 +if test "$enable_sse" = "yes"; then
   4.143 +	AC_DEFINE([HAVE_SSE], [1], [Define to 1 to enable sse optimizations.])
   4.144 +	AC_DEFINE([HAVE_MMX], [1], [Define to 1 to enable mmx optimizations.])
   4.145 +	AC_DEFINE([HAVE_MMX2], [1], [Define to 1 to enable mmx2 optimizations.])
   4.146 +	ARCH_SUBDIR=x86
   4.147 +fi
   4.148 +
   4.149 +AC_ARG_ENABLE([altivec], AS_HELP_STRING([--enable-altivec], [Enable altivec optimizations]))
   4.150 +if test "$enable_altivec" = "yes"; then
   4.151 +	AC_DEFINE([HAVE_ALTIVEC], [1], [Define to 1 to enable altivec optimizations.])
   4.152 +	ARCH_SUBDIR="$ARCH_SUBDIR ppc"
   4.153 +	TMPCLAGS=$CFLAGS
   4.154 +	CFLAGS="$CFLAGS -maltivec"
   4.155 +	AC_CHECK_HEADERS(altivec.h)
   4.156 +	CFLAGS=$TMPCLAGS
   4.157 +fi
   4.158 +
   4.159 +AC_ARG_ENABLE([neon], AS_HELP_STRING([--enable-neon], [Enable neon optimizations]))
   4.160 +if test "$enable_neon" = "yes"; then
   4.161 +	AC_DEFINE([HAVE_NEON], [1], [Define to 1 to enable neon optimizations.])
   4.162 +	ARCH_SUBDIR=arm
   4.163 +fi
   4.164 +
   4.165 +AM_CONDITIONAL([HAVE_ARCH_SUBDIR], [test "$ARCH_SUBDIR" != ""])
   4.166 +AC_SUBST([ARCH_SUBDIR])
   4.167 +
   4.168 +AC_DEFINE([HAVE_NEON], [0], [Define to 1 to enable neon optimizations.])
   4.169 +
   4.170 +AC_CONFIG_HEADER([config.h])
   4.171 +
   4.172 +AC_CONFIG_FILES([Makefile libavutil/Makefile libavcodec/Makefile libavcodec/x86/Makefile libavcodec/ppc/Makefile libavcodec/cell/Makefile])
   4.173 +
   4.174 +AC_OUTPUT

     5.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     5.2 +++ b/ffmpeg_smp/h264dec/h264dec.c	Mon Aug 27 12:09:56 2012 +0200
     5.3 @@ -0,0 +1,288 @@
     5.4 +/*
     5.5 +* H264 decoder main
     5.6 +*/
     5.7 +
     5.8 +#include "config.h"
     5.9 +#include "libavcodec/h264.h"
    5.10 +
    5.11 +#include <string.h>
    5.12 +#include <stdlib.h>
    5.13 +#include <errno.h>
    5.14 +#include <signal.h>
    5.15 +#include <unistd.h>
    5.16 +#include <getopt.h>
    5.17 +#include <fcntl.h>
    5.18 +
    5.19 +#include <sys/types.h>
    5.20 +#include <sys/time.h>
    5.21 +#include <sys/resource.h>
    5.22 +#include <time.h>
    5.23 +
    5.24 +#include <assert.h>
    5.25 +
    5.26 +
    5.27 +static const char program_name[] = "h264dec";
    5.28 +static const int program_birth_year = 2010;
    5.29 +
    5.30 +static const char *file_name;
    5.31 +static int ifile, ofile;
    5.32 +static int no_arch =0;
    5.33 +static int parallel = 1;
    5.34 +static int frame_width  = 0;
    5.35 +static int frame_height = 0;
    5.36 +
    5.37 +static void av_exit(int ret)
    5.38 +{
    5.39 +    //do some free calls
    5.40 +#undef exit
    5.41 +    exit(ret);
    5.42 +}
    5.43 +
    5.44 +static void opt_input_file(const char *filename)
    5.45 +{
    5.46 +    /* open the input file */
    5.47 +    ifile = open(filename, O_RDONLY, 0666);
    5.48 +    if (ifile < 0){
    5.49 +        fprintf(stderr, "Failed to open %s\n", filename);
    5.50 +        av_exit(-1);
    5.51 +    }
    5.52 +
    5.53 +    //parse first frame to get resolution (other information available but not used)
    5.54 +    H264Slice slice;
    5.55 +    PictureInfo pi;
    5.56 +    GetBitContext gb = {0,};
    5.57 +    ParserContext *pc;
    5.58 +    NalContext *nc;
    5.59 +
    5.60 +    pc = get_parse_context(ifile);
    5.61 +    nc = get_nal_context(0, 0);
    5.62 +
    5.63 +    memset(&slice, 0, sizeof(H264Slice));
    5.64 +    slice.current_picture_info=&pi;
    5.65 +
    5.66 +    av_read_frame_internal(pc, &gb);
    5.67 +    decode_nal_units(nc, &slice, &gb);
    5.68 +
    5.69 +    frame_width = nc->width;
    5.70 +    frame_height= nc->height;
    5.71 +
    5.72 +    //clean up
    5.73 +    av_freep(&gb.raw);
    5.74 +    if (gb.rbsp)
    5.75 +        av_freep(&gb.rbsp);
    5.76 +    free_parse_context(pc);
    5.77 +    free_nal_context(nc);
    5.78 +
    5.79 +    //rewind file
    5.80 +    int offset;
    5.81 +    if ( (offset=lseek(ifile, 0, SEEK_SET)) ){
    5.82 +        fprintf(stderr, "Rewind input file %s failed at offset %d\n", filename, offset);
    5.83 +    }
    5.84 +
    5.85 +}
    5.86 +
    5.87 +static void opt_output_file(const char *filename)
    5.88 +{
    5.89 +    if (filename){
    5.90 +        if (!strcmp(filename, "-"))
    5.91 +            filename = "pipe:";
    5.92 +
    5.93 +        ofile = open(filename, O_CREAT | O_TRUNC | O_WRONLY, 0666);
    5.94 +    }else{
    5.95 +        ofile =0;
    5.96 +    }
    5.97 +}
    5.98 +
    5.99 +static void show_usage(void)
   5.100 +{
   5.101 +    printf("usage: ffmpeg [options] -i infile }...\n");
   5.102 +    printf("\n");
   5.103 +}
   5.104 +
   5.105 +static struct option long_options[] = {
   5.106 +    {"static-sched", 0, 0, 0},
   5.107 +    {"static-mbd", 0, 0, 0},
   5.108 +    {"numamap", 0, 0, 0},
   5.109 +    {"no-mbd", 0, 0, 0},
   5.110 +    {"static-3d", 0, 0, 0},
   5.111 +    {"slice-bufs", 1, 0, 0},
   5.112 +    {"smt", 0, 0, 0},
   5.113 +    {"noarch", 0, 0, 'a'},
   5.114 +    {"display", 0, 0, 'd'},
   5.115 +    {"fullscreen", 0, 0, 'f'},
   5.116 +    {"numframes", 1, 0, 'n'},
   5.117 +    {"use-ppe-ed", 1, 0, 'p'},
   5.118 +    {"sequential", 0, 0, 's'},
   5.119 +    {"threads", 1, 0, 't'},
   5.120 +    {"verbose", 1, 0, 'v'},
   5.121 +    {"wave-order", 1, 0, 'w'},
   5.122 +    {"smb-size", 1, 0, 'z'},
   5.123 +    {"pipe-bufs", 1, 0, 'e'},
   5.124 +    {0, 0, 0, 0}
   5.125 +};
   5.126 +
   5.127 +static h264_options cli_opts;
   5.128 +static void parse_cmd(int argc, char **argv)
   5.129 +{
   5.130 +    int c;
   5.131 +    int digit_optind = 0;
   5.132 +    int option_index = 0;
   5.133 +    char ofile_name[1024];
   5.134 +    extern char *optarg;
   5.135 +    extern int optind, optopt;
   5.136 +
   5.137 +    cli_opts.statsched =0;
   5.138 +    cli_opts.numamap =0;
   5.139 +    cli_opts.statmbd =0;
   5.140 +    cli_opts.no_mbd= 0;
   5.141 +    cli_opts.numframes = INT_MAX;
   5.142 +    cli_opts.display=0;
   5.143 +    cli_opts.fullscreen=0;
   5.144 +    cli_opts.verbose=0;
   5.145 +    cli_opts.ppe_ed=0;
   5.146 +    cli_opts.profile=0;
   5.147 +    cli_opts.threads = 1;
   5.148 +    cli_opts.smb_size[0] = cli_opts.smb_size[1] = 1;
   5.149 +    cli_opts.wave_order=0;
   5.150 +    cli_opts.static_3d=0;
   5.151 +    cli_opts.pipe_bufs=8;
   5.152 +    cli_opts.slice_bufs=1;
   5.153 +    cli_opts.smt= 0;
   5.154 +    while ((c = getopt_long(argc, argv, "ade:fi:n:o:p:st:vwz:", long_options, &option_index)) != -1 ){
   5.155 +        int this_option_optind = optind ? optind : 1;
   5.156 +
   5.157 +        switch (c){
   5.158 +            case 0:
   5.159 +                if (option_index==0){
   5.160 +                    cli_opts.statsched=1;
   5.161 +                }else if (option_index==1){
   5.162 +                    cli_opts.statmbd= 1;
   5.163 +                }else if (option_index==2){
   5.164 +                    cli_opts.numamap= 1;
   5.165 +                }else if (option_index==3){
   5.166 +                    cli_opts.no_mbd= 1;
   5.167 +                }else if (option_index==4){
   5.168 +                    cli_opts.static_3d= 1;
   5.169 +                }else if (option_index==5){
   5.170 +                    cli_opts.slice_bufs= (unsigned) atoi(optarg);
   5.171 +                }else if (option_index==6){
   5.172 +                    cli_opts.smt= 1;
   5.173 +                }
   5.174 +                break;
   5.175 +            case '0':
   5.176 +            case '1':
   5.177 +            case '2':
   5.178 +                if (digit_optind != 0 && digit_optind != this_option_optind)
   5.179 +                    printf("digits occur in two different argv-elements.\n");
   5.180 +                digit_optind = this_option_optind;
   5.181 +                printf("option %c\n", c);
   5.182 +                break;
   5.183 +            case 'a':
   5.184 +                no_arch=1;
   5.185 +                break;
   5.186 +            case 'd':
   5.187 +                cli_opts.display=1;
   5.188 +                break;
   5.189 +            case 'f':
   5.190 +                cli_opts.fullscreen=1;
   5.191 +                break;
   5.192 +            case 'i':
   5.193 +                file_name = (const char *)optarg;
   5.194 +                opt_input_file(file_name);
   5.195 +                break;
   5.196 +            case 'n':
   5.197 +                cli_opts.numframes = (unsigned) atoi(optarg);
   5.198 +                break;
   5.199 +            case 'o':
   5.200 +                strcpy(ofile_name, optarg);
   5.201 +                opt_output_file(ofile_name);
   5.202 +                break;
   5.203 +            case 'p':
   5.204 +                cli_opts.profile = (unsigned) atoi(optarg);
   5.205 +                break;
   5.206 +            case 's':
   5.207 +                cli_opts.threads = 0;
   5.208 +                parallel = 0;
   5.209 +                break;
   5.210 +            case 't':
   5.211 +                cli_opts.threads = atoi(optarg);
   5.212 +                if (cli_opts.threads<=0){
   5.213 +                    fprintf(stderr, "Option -%c requires thread numbers > 0\n", c);
   5.214 +                    av_exit(-1);
   5.215 +                }
   5.216 +                break;
   5.217 +            case 'v':
   5.218 +                cli_opts.verbose = 1;
   5.219 +                break;
   5.220 +            case 'w':
   5.221 +                cli_opts.wave_order = 1;
   5.222 +                break;
   5.223 +            case 'z': // only useful in ompss
   5.224 +                if (argc < optind +1){
   5.225 +                    fprintf(stderr, "Option -%c (--smb-size) requires 2 arguments\n", c);
   5.226 +                    av_exit(-1);
   5.227 +                }
   5.228 +                optind--;
   5.229 +                for (int i=0; i<2; i++){
   5.230 +                    cli_opts.smb_size[i] = atoi(argv[optind++]);
   5.231 +                    if (!(cli_opts.smb_size > 0)){
   5.232 +                        fprintf(stderr, "Option -%c (--smb-size) requires dimensions > 0\n", c);
   5.233 +                        av_exit(-1);
   5.234 +                    }
   5.235 +                }
   5.236 +                break;
   5.237 +            case 'e':
   5.238 +                cli_opts.pipe_bufs = atoi(optarg);
   5.239 +                break;
   5.240 +            case ':':
   5.241 +                fprintf(stderr, "Option -%c requires an operand\n", optopt);
   5.242 +                av_exit(-1);
   5.243 +                break;
   5.244 +            case '?':
   5.245 +                fprintf(stderr, "Unrecognized option: -%c\n", optopt);
   5.246 +                av_exit(-1);
   5.247 +                break;
   5.248 +        }
   5.249 +    }
   5.250 +
   5.251 +}
   5.252 +
   5.253 +int main(int argc, char **argv)
   5.254 +{
   5.255 +    /* parse options */
   5.256 +    parse_cmd(argc, argv);
   5.257 +
   5.258 +    if(!ifile ) {
   5.259 +        show_usage();
   5.260 +        av_exit(1);
   5.261 +    }
   5.262 +
   5.263 +    H264Context *h = get_h264dec_context(file_name, ifile, ofile, frame_width, frame_height, &cli_opts);
   5.264 +#if OMPSS
   5.265 +    if (h264_decode_ompss( h ) < 0)
   5.266 +        av_exit(-1);
   5.267 +#else
   5.268 +    if (parallel){
   5.269 +        if (ARCH_CELL && !no_arch){
   5.270 +            if (h264_decode_cell( h ) < 0)
   5.271 +                av_exit(-1);
   5.272 +        }else{
   5.273 +            if (h264_decode_pthread( h ) < 0)
   5.274 +                av_exit(1);
   5.275 +        }
   5.276 +    }else{
   5.277 +        if (ARCH_CELL && !no_arch){
   5.278 +            if (h264_decode_cell_seq( h ) < 0)
   5.279 +                av_exit(1);
   5.280 +        }else{
   5.281 +            if (h264_decode_seq( h ) < 0)
   5.282 +                av_exit(1);
   5.283 +        }
   5.284 +    }
   5.285 +#endif
   5.286 +    free_h264dec_context(h);
   5.287 +    close(ifile);
   5.288 +    close(ofile);
   5.289 +
   5.290 +    return 0;
   5.291 +}

     6.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     6.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/aac.h	Mon Aug 27 12:09:56 2012 +0200
     6.3 @@ -0,0 +1,137 @@
     6.4 +/*
     6.5 + * Copyright (c) 2010 Mans Rullgard <mans@mansr.com>
     6.6 + *
     6.7 + * This file is part of FFmpeg.
     6.8 + *
     6.9 + * FFmpeg is free software; you can redistribute it and/or
    6.10 + * modify it under the terms of the GNU Lesser General Public
    6.11 + * License as published by the Free Software Foundation; either
    6.12 + * version 2.1 of the License, or (at your option) any later version.
    6.13 + *
    6.14 + * FFmpeg is distributed in the hope that it will be useful,
    6.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
    6.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    6.17 + * Lesser General Public License for more details.
    6.18 + *
    6.19 + * You should have received a copy of the GNU Lesser General Public
    6.20 + * License along with FFmpeg; if not, write to the Free Software
    6.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
    6.22 + */
    6.23 +
    6.24 +#ifndef AVCODEC_ARM_AAC_H
    6.25 +#define AVCODEC_ARM_AAC_H
    6.26 +
    6.27 +#include "config.h"
    6.28 +
    6.29 +#if HAVE_NEON && HAVE_INLINE_ASM
    6.30 +
    6.31 +#define VMUL2 VMUL2
    6.32 +static inline float *VMUL2(float *dst, const float *v, unsigned idx,
    6.33 +                           const float *scale)
    6.34 +{
    6.35 +    unsigned v0, v1;
    6.36 +    __asm__ volatile ("ubfx     %0,  %4,  #0, #4      \n\t"
    6.37 +                      "ubfx     %1,  %4,  #4, #4      \n\t"
    6.38 +                      "ldr      %0,  [%3, %0, lsl #2] \n\t"
    6.39 +                      "ldr      %1,  [%3, %1, lsl #2] \n\t"
    6.40 +                      "vld1.32  {d1[]},   [%5,:32]    \n\t"
    6.41 +                      "vmov     d0,  %0,  %1          \n\t"
    6.42 +                      "vmul.f32 d0,  d0,  d1          \n\t"
    6.43 +                      "vst1.32  {d0},     [%2,:64]!   \n\t"
    6.44 +                      : "=&r"(v0), "=&r"(v1), "+r"(dst)
    6.45 +                      : "r"(v), "r"(idx), "r"(scale)
    6.46 +                      : "d0", "d1");
    6.47 +    return dst;
    6.48 +}
    6.49 +
    6.50 +#define VMUL4 VMUL4
    6.51 +static inline float *VMUL4(float *dst, const float *v, unsigned idx,
    6.52 +                           const float *scale)
    6.53 +{
    6.54 +    unsigned v0, v1, v2, v3;
    6.55 +    __asm__ volatile ("ubfx     %0,  %6,  #0, #2      \n\t"
    6.56 +                      "ubfx     %1,  %6,  #2, #2      \n\t"
    6.57 +                      "ldr      %0,  [%5, %0, lsl #2] \n\t"
    6.58 +                      "ubfx     %2,  %6,  #4, #2      \n\t"
    6.59 +                      "ldr      %1,  [%5, %1, lsl #2] \n\t"
    6.60 +                      "ubfx     %3,  %6,  #6, #2      \n\t"
    6.61 +                      "ldr      %2,  [%5, %2, lsl #2] \n\t"
    6.62 +                      "vmov     d0,  %0,  %1          \n\t"
    6.63 +                      "ldr      %3,  [%5, %3, lsl #2] \n\t"
    6.64 +                      "vld1.32  {d2[],d3[]},[%7,:32]  \n\t"
    6.65 +                      "vmov     d1,  %2,  %3          \n\t"
    6.66 +                      "vmul.f32 q0,  q0,  q1          \n\t"
    6.67 +                      "vst1.32  {q0},     [%4,:128]!  \n\t"
    6.68 +                      : "=&r"(v0), "=&r"(v1), "=&r"(v2), "=&r"(v3), "+r"(dst)
    6.69 +                      : "r"(v), "r"(idx), "r"(scale)
    6.70 +                      : "d0", "d1", "d2", "d3");
    6.71 +    return dst;
    6.72 +}
    6.73 +
    6.74 +#define VMUL2S VMUL2S
    6.75 +static inline float *VMUL2S(float *dst, const float *v, unsigned idx,
    6.76 +                            unsigned sign, const float *scale)
    6.77 +{
    6.78 +    unsigned v0, v1, v2, v3;
    6.79 +    __asm__ volatile ("ubfx     %0,  %6,  #0, #4      \n\t"
    6.80 +                      "ubfx     %1,  %6,  #4, #4      \n\t"
    6.81 +                      "ldr      %0,  [%5, %0, lsl #2] \n\t"
    6.82 +                      "lsl      %2,  %8,  #30         \n\t"
    6.83 +                      "ldr      %1,  [%5, %1, lsl #2] \n\t"
    6.84 +                      "lsl      %3,  %8,  #31         \n\t"
    6.85 +                      "vmov     d0,  %0,  %1          \n\t"
    6.86 +                      "bic      %2,  %2,  #1<<30      \n\t"
    6.87 +                      "vld1.32  {d1[]},   [%7,:32]    \n\t"
    6.88 +                      "vmov     d2,  %2,  %3          \n\t"
    6.89 +                      "veor     d0,  d0,  d2          \n\t"
    6.90 +                      "vmul.f32 d0,  d0,  d1          \n\t"
    6.91 +                      "vst1.32  {d0},     [%4,:64]!   \n\t"
    6.92 +                      : "=&r"(v0), "=&r"(v1), "=&r"(v2), "=&r"(v3), "+r"(dst)
    6.93 +                      : "r"(v), "r"(idx), "r"(scale), "r"(sign)
    6.94 +                      : "d0", "d1", "d2");
    6.95 +    return dst;
    6.96 +}
    6.97 +
    6.98 +#define VMUL4S VMUL4S
    6.99 +static inline float *VMUL4S(float *dst, const float *v, unsigned idx,
   6.100 +                            unsigned sign, const float *scale)
   6.101 +{
   6.102 +    unsigned v0, v1, v2, v3, nz;
   6.103 +    __asm__ volatile ("vld1.32  {d2[],d3[]},[%9,:32]  \n\t"
   6.104 +                      "ubfx     %0,  %8,  #0, #2      \n\t"
   6.105 +                      "ubfx     %1,  %8,  #2, #2      \n\t"
   6.106 +                      "ldr      %0,  [%7, %0, lsl #2] \n\t"
   6.107 +                      "ubfx     %2,  %8,  #4, #2      \n\t"
   6.108 +                      "ldr      %1,  [%7, %1, lsl #2] \n\t"
   6.109 +                      "ubfx     %3,  %8,  #6, #2      \n\t"
   6.110 +                      "ldr      %2,  [%7, %2, lsl #2] \n\t"
   6.111 +                      "vmov     d0,  %0,  %1          \n\t"
   6.112 +                      "ldr      %3,  [%7, %3, lsl #2] \n\t"
   6.113 +                      "lsr      %6,  %8,  #12         \n\t"
   6.114 +                      "rbit     %6,  %6               \n\t"
   6.115 +                      "vmov     d1,  %2,  %3          \n\t"
   6.116 +                      "lsls     %6,  %6,  #1          \n\t"
   6.117 +                      "and      %0,  %5,  #1<<31      \n\t"
   6.118 +                      "lslcs    %5,  %5,  #1          \n\t"
   6.119 +                      "lsls     %6,  %6,  #1          \n\t"
   6.120 +                      "and      %1,  %5,  #1<<31      \n\t"
   6.121 +                      "lslcs    %5,  %5,  #1          \n\t"
   6.122 +                      "lsls     %6,  %6,  #1          \n\t"
   6.123 +                      "and      %2,  %5,  #1<<31      \n\t"
   6.124 +                      "lslcs    %5,  %5,  #1          \n\t"
   6.125 +                      "vmov     d4,  %0,  %1          \n\t"
   6.126 +                      "and      %3,  %5,  #1<<31      \n\t"
   6.127 +                      "vmov     d5,  %2,  %3          \n\t"
   6.128 +                      "veor     q0,  q0,  q2          \n\t"
   6.129 +                      "vmul.f32 q0,  q0,  q1          \n\t"
   6.130 +                      "vst1.32  {q0},     [%4,:128]!  \n\t"
   6.131 +                      : "=&r"(v0), "=&r"(v1), "=&r"(v2), "=&r"(v3), "+r"(dst),
   6.132 +                        "+r"(sign), "=r"(nz)
   6.133 +                      : "r"(v), "r"(idx), "r"(scale)
   6.134 +                      : "d0", "d1", "d2", "d3", "d4", "d5");
   6.135 +    return dst;
   6.136 +}
   6.137 +
   6.138 +#endif /* HAVE_NEON && HAVE_INLINE_ASM */
   6.139 +
   6.140 +#endif /* AVCODEC_ARM_AAC_H */

     7.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     7.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/asm.S	Mon Aug 27 12:09:56 2012 +0200
     7.3 @@ -0,0 +1,72 @@
     7.4 +/*
     7.5 + * Copyright (c) 2008 Mans Rullgard <mans@mansr.com>
     7.6 + *
     7.7 + * This file is part of FFmpeg.
     7.8 + *
     7.9 + * FFmpeg is free software; you can redistribute it and/or
    7.10 + * modify it under the terms of the GNU Lesser General Public
    7.11 + * License as published by the Free Software Foundation; either
    7.12 + * version 2.1 of the License, or (at your option) any later version.
    7.13 + *
    7.14 + * FFmpeg is distributed in the hope that it will be useful,
    7.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
    7.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    7.17 + * Lesser General Public License for more details.
    7.18 + *
    7.19 + * You should have received a copy of the GNU Lesser General Public
    7.20 + * License along with FFmpeg; if not, write to the Free Software
    7.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
    7.22 + */
    7.23 +
    7.24 +#include "config.h"
    7.25 +
    7.26 +#ifdef __ELF__
    7.27 +#   define ELF
    7.28 +#else
    7.29 +#   define ELF @
    7.30 +#endif
    7.31 +
    7.32 +        .macro require8, val=1
    7.33 +ELF     .eabi_attribute 24, \val
    7.34 +        .endm
    7.35 +
    7.36 +        .macro preserve8, val=1
    7.37 +ELF     .eabi_attribute 25, \val
    7.38 +        .endm
    7.39 +
    7.40 +        .macro function name, export=0
    7.41 +        .macro endfunc
    7.42 +ELF     .size   \name, . - \name
    7.43 +        .endfunc
    7.44 +        .purgem endfunc
    7.45 +        .endm
    7.46 +.if \export
    7.47 +        .global EXTERN_ASM\name
    7.48 +EXTERN_ASM\name:
    7.49 +.endif
    7.50 +ELF     .type   \name, %function
    7.51 +        .func   \name
    7.52 +\name:
    7.53 +        .endm
    7.54 +
    7.55 +        .macro movrel rd, val
    7.56 +#if HAVE_ARMV6T2 && !CONFIG_PIC
    7.57 +        movw            \rd, #:lower16:\val
    7.58 +        movt            \rd, #:upper16:\val
    7.59 +#else
    7.60 +        ldr             \rd, =\val
    7.61 +#endif
    7.62 +        .endm
    7.63 +
    7.64 +#if HAVE_VFP_ARGS
    7.65 +        .eabi_attribute 28, 1
    7.66 +#   define VFP
    7.67 +#   define NOVFP @
    7.68 +#else
    7.69 +#   define VFP   @
    7.70 +#   define NOVFP
    7.71 +#endif
    7.72 +
    7.73 +#define GLUE(a, b) a ## b
    7.74 +#define JOIN(a, b) GLUE(a, b)
    7.75 +#define X(s) JOIN(EXTERN_ASM, s)

     8.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     8.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dcadsp_init_arm.c	Mon Aug 27 12:09:56 2012 +0200
     8.3 @@ -0,0 +1,32 @@
     8.4 +/*
     8.5 + * Copyright (c) 2010 Mans Rullgard <mans@mansr.com>
     8.6 + *
     8.7 + * This file is part of FFmpeg.
     8.8 + *
     8.9 + * FFmpeg is free software; you can redistribute it and/or
    8.10 + * modify it under the terms of the GNU Lesser General Public
    8.11 + * License as published by the Free Software Foundation; either
    8.12 + * version 2.1 of the License, or (at your option) any later version.
    8.13 + *
    8.14 + * FFmpeg is distributed in the hope that it will be useful,
    8.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
    8.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    8.17 + * Lesser General Public License for more details.
    8.18 + *
    8.19 + * You should have received a copy of the GNU Lesser General Public
    8.20 + * License along with FFmpeg; if not, write to the Free Software
    8.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
    8.22 + */
    8.23 +
    8.24 +#include "config.h"
    8.25 +#include "libavutil/attributes.h"
    8.26 +#include "libavcodec/dcadsp.h"
    8.27 +
    8.28 +void ff_dca_lfe_fir_neon(float *out, const float *in, const float *coefs,
    8.29 +                         int decifactor, float scale, float bias);
    8.30 +
    8.31 +void av_cold ff_dcadsp_init_arm(DCADSPContext *s)
    8.32 +{
    8.33 +    if (HAVE_NEON)
    8.34 +        s->lfe_fir = ff_dca_lfe_fir_neon;
    8.35 +}

     9.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     9.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dcadsp_neon.S	Mon Aug 27 12:09:56 2012 +0200
     9.3 @@ -0,0 +1,61 @@
     9.4 +/*
     9.5 + * Copyright (c) 2010 Mans Rullgard <mans@mansr.com>
     9.6 + *
     9.7 + * This file is part of FFmpeg.
     9.8 + *
     9.9 + * FFmpeg is free software; you can redistribute it and/or
    9.10 + * modify it under the terms of the GNU Lesser General Public
    9.11 + * License as published by the Free Software Foundation; either
    9.12 + * version 2.1 of the License, or (at your option) any later version.
    9.13 + *
    9.14 + * FFmpeg is distributed in the hope that it will be useful,
    9.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
    9.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    9.17 + * Lesser General Public License for more details.
    9.18 + *
    9.19 + * You should have received a copy of the GNU Lesser General Public
    9.20 + * License along with FFmpeg; if not, write to the Free Software
    9.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
    9.22 + */
    9.23 +
    9.24 +#include "asm.S"
    9.25 +
    9.26 +function ff_dca_lfe_fir_neon, export=1
    9.27 +        push            {r4-r6,lr}
    9.28 +
    9.29 +        add             r4,  r0,  r3,  lsl #2   @ out2
    9.30 +        add             r5,  r2,  #256*4-16     @ cf1
    9.31 +        sub             r1,  r1,  #12
    9.32 +        cmp             r3,  #32
    9.33 +        moveq           r6,  #256/32
    9.34 +        movne           r6,  #256/64
    9.35 +NOVFP   vldr            d0,  [sp, #16]          @ scale, bias
    9.36 +        mov             lr,  #-16
    9.37 +1:
    9.38 +        vmov.f32        q2,  #0.0               @ v0
    9.39 +        vmov.f32        q3,  #0.0               @ v1
    9.40 +        mov             r12, r6
    9.41 +2:
    9.42 +        vld1.32         {q8},     [r2,:128]!    @ cf0
    9.43 +        vld1.32         {q9},     [r5,:128], lr @ cf1
    9.44 +        vld1.32         {q1},     [r1], lr      @ in
    9.45 +        subs            r12, r12, #4
    9.46 +        vrev64.32       q10, q8
    9.47 +        vmla.f32        q3,  q1,  q9
    9.48 +        vmla.f32        d4,  d2,  d21
    9.49 +        vmla.f32        d5,  d3,  d20
    9.50 +        bne             2b
    9.51 +
    9.52 +        add             r1,  r1,  r6,  lsl #2
    9.53 +        subs            r3,  r3,  #1
    9.54 +        vadd.f32        d4,  d4,  d5
    9.55 +        vadd.f32        d6,  d6,  d7
    9.56 +        vpadd.f32       d4,  d4,  d6
    9.57 +        vdup.32         d5,  d0[1]
    9.58 +        vmla.f32        d5,  d4,  d0[0]
    9.59 +        vst1.32         {d5[0]},  [r0,:32]!
    9.60 +        vst1.32         {d5[1]},  [r4,:32]!
    9.61 +        bne             1b
    9.62 +
    9.63 +        pop             {r4-r6,pc}
    9.64 +endfunc

    10.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
    10.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dsputil_arm.S	Mon Aug 27 12:09:56 2012 +0200
    10.3 @@ -0,0 +1,712 @@
    10.4 +@
    10.5 +@ ARMv4 optimized DSP utils
    10.6 +@ Copyright (c) 2004 AGAWA Koji <i (AT) atty (DOT) jp>
    10.7 +@
    10.8 +@ This file is part of FFmpeg.
    10.9 +@
   10.10 +@ FFmpeg is free software; you can redistribute it and/or
   10.11 +@ modify it under the terms of the GNU Lesser General Public
   10.12 +@ License as published by the Free Software Foundation; either
   10.13 +@ version 2.1 of the License, or (at your option) any later version.
   10.14 +@
   10.15 +@ FFmpeg is distributed in the hope that it will be useful,
   10.16 +@ but WITHOUT ANY WARRANTY; without even the implied warranty of
   10.17 +@ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   10.18 +@ Lesser General Public License for more details.
   10.19 +@
   10.20 +@ You should have received a copy of the GNU Lesser General Public
   10.21 +@ License along with FFmpeg; if not, write to the Free Software
   10.22 +@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
   10.23 +@
   10.24 +
   10.25 +#include "config.h"
   10.26 +#include "asm.S"
   10.27 +
   10.28 +        preserve8
   10.29 +
   10.30 +#if !HAVE_PLD
   10.31 +.macro pld reg
   10.32 +.endm
   10.33 +#endif
   10.34 +
   10.35 +#if HAVE_ARMV5TE
   10.36 +function ff_prefetch_arm, export=1
   10.37 +        subs            r2,  r2,  #1
   10.38 +        pld             [r0]
   10.39 +        add             r0,  r0,  r1
   10.40 +        bne             ff_prefetch_arm
   10.41 +        bx              lr
   10.42 +endfunc
   10.43 +#endif
   10.44 +
   10.45 +.macro  ALIGN_QWORD_D shift, Rd0, Rd1, Rd2, Rd3, Rn0, Rn1, Rn2, Rn3, Rn4
   10.46 +        mov             \Rd0, \Rn0, lsr #(\shift * 8)
   10.47 +        mov             \Rd1, \Rn1, lsr #(\shift * 8)
   10.48 +        mov             \Rd2, \Rn2, lsr #(\shift * 8)
   10.49 +        mov             \Rd3, \Rn3, lsr #(\shift * 8)
   10.50 +        orr             \Rd0, \Rd0, \Rn1, lsl #(32 - \shift * 8)
   10.51 +        orr             \Rd1, \Rd1, \Rn2, lsl #(32 - \shift * 8)
   10.52 +        orr             \Rd2, \Rd2, \Rn3, lsl #(32 - \shift * 8)
   10.53 +        orr             \Rd3, \Rd3, \Rn4, lsl #(32 - \shift * 8)
   10.54 +.endm
   10.55 +.macro  ALIGN_DWORD shift, R0, R1, R2
   10.56 +        mov             \R0, \R0, lsr #(\shift * 8)
   10.57 +        orr             \R0, \R0, \R1, lsl #(32 - \shift * 8)
   10.58 +        mov             \R1, \R1, lsr #(\shift * 8)
   10.59 +        orr             \R1, \R1, \R2, lsl #(32 - \shift * 8)
   10.60 +.endm
   10.61 +.macro  ALIGN_DWORD_D shift, Rdst0, Rdst1, Rsrc0, Rsrc1, Rsrc2
   10.62 +        mov             \Rdst0, \Rsrc0, lsr #(\shift * 8)
   10.63 +        mov             \Rdst1, \Rsrc1, lsr #(\shift * 8)
   10.64 +        orr             \Rdst0, \Rdst0, \Rsrc1, lsl #(32 - (\shift * 8))
   10.65 +        orr             \Rdst1, \Rdst1, \Rsrc2, lsl #(32 - (\shift * 8))
   10.66 +.endm
   10.67 +
   10.68 +.macro  RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask
   10.69 +        @ Rd = (Rn | Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1)
   10.70 +        @ Rmask = 0xFEFEFEFE
   10.71 +        @ Rn = destroy
   10.72 +        eor             \Rd0, \Rn0, \Rm0
   10.73 +        eor             \Rd1, \Rn1, \Rm1
   10.74 +        orr             \Rn0, \Rn0, \Rm0
   10.75 +        orr             \Rn1, \Rn1, \Rm1
   10.76 +        and             \Rd0, \Rd0, \Rmask
   10.77 +        and             \Rd1, \Rd1, \Rmask
   10.78 +        sub             \Rd0, \Rn0, \Rd0, lsr #1
   10.79 +        sub             \Rd1, \Rn1, \Rd1, lsr #1
   10.80 +.endm
   10.81 +
   10.82 +.macro  NO_RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask
   10.83 +        @ Rd = (Rn & Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1)
   10.84 +        @ Rmask = 0xFEFEFEFE
   10.85 +        @ Rn = destroy
   10.86 +        eor             \Rd0, \Rn0, \Rm0
   10.87 +        eor             \Rd1, \Rn1, \Rm1
   10.88 +        and             \Rn0, \Rn0, \Rm0
   10.89 +        and             \Rn1, \Rn1, \Rm1
   10.90 +        and             \Rd0, \Rd0, \Rmask
   10.91 +        and             \Rd1, \Rd1, \Rmask
   10.92 +        add             \Rd0, \Rn0, \Rd0, lsr #1
   10.93 +        add             \Rd1, \Rn1, \Rd1, lsr #1
   10.94 +.endm
   10.95 +
   10.96 +.macro  JMP_ALIGN tmp, reg
   10.97 +        ands            \tmp, \reg, #3
   10.98 +        bic             \reg, \reg, #3
   10.99 +        beq             1f
  10.100 +        subs            \tmp, \tmp, #1
  10.101 +        beq             2f
  10.102 +        subs            \tmp, \tmp, #1
  10.103 +        beq             3f
  10.104 +        b    4f
  10.105 +.endm
  10.106 +
  10.107 +@ ----------------------------------------------------------------
  10.108 +        .align 5
  10.109 +function ff_put_pixels16_arm, export=1
  10.110 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.111 +        @ block = word aligned, pixles = unaligned
  10.112 +        pld             [r1]
  10.113 +        push            {r4-r11, lr}
  10.114 +        JMP_ALIGN       r5,  r1
  10.115 +1:
  10.116 +        ldm             r1,  {r4-r7}
  10.117 +        add             r1,  r1,  r2
  10.118 +        stm             r0,  {r4-r7}
  10.119 +        pld             [r1]
  10.120 +        subs            r3,  r3,  #1
  10.121 +        add             r0,  r0,  r2
  10.122 +        bne             1b
  10.123 +        pop             {r4-r11, pc}
  10.124 +        .align 5
  10.125 +2:
  10.126 +        ldm             r1,  {r4-r8}
  10.127 +        add             r1,  r1,  r2
  10.128 +        ALIGN_QWORD_D   1,   r9,  r10, r11, r12, r4,  r5,  r6,  r7,  r8
  10.129 +        pld             [r1]
  10.130 +        subs            r3,  r3,  #1
  10.131 +        stm             r0,  {r9-r12}
  10.132 +        add             r0,  r0,  r2
  10.133 +        bne             2b
  10.134 +        pop             {r4-r11, pc}
  10.135 +        .align 5
  10.136 +3:
  10.137 +        ldm             r1,  {r4-r8}
  10.138 +        add             r1,  r1,  r2
  10.139 +        ALIGN_QWORD_D   2,   r9,  r10, r11, r12, r4,  r5,  r6,  r7,  r8
  10.140 +        pld             [r1]
  10.141 +        subs            r3,  r3,  #1
  10.142 +        stm             r0,  {r9-r12}
  10.143 +        add             r0,  r0,  r2
  10.144 +        bne             3b
  10.145 +        pop             {r4-r11, pc}
  10.146 +        .align 5
  10.147 +4:
  10.148 +        ldm             r1,  {r4-r8}
  10.149 +        add             r1,  r1,  r2
  10.150 +        ALIGN_QWORD_D   3,   r9,  r10, r11, r12, r4,  r5,  r6,  r7,  r8
  10.151 +        pld             [r1]
  10.152 +        subs            r3,  r3,  #1
  10.153 +        stm             r0,  {r9-r12}
  10.154 +        add             r0,  r0,  r2
  10.155 +        bne             4b
  10.156 +        pop             {r4-r11,pc}
  10.157 +endfunc
  10.158 +
  10.159 +@ ----------------------------------------------------------------
  10.160 +        .align 5
  10.161 +function ff_put_pixels8_arm, export=1
  10.162 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.163 +        @ block = word aligned, pixles = unaligned
  10.164 +        pld             [r1]
  10.165 +        push            {r4-r5,lr}
  10.166 +        JMP_ALIGN       r5,  r1
  10.167 +1:
  10.168 +        ldm             r1,  {r4-r5}
  10.169 +        add             r1,  r1,  r2
  10.170 +        subs            r3,  r3,  #1
  10.171 +        pld             [r1]
  10.172 +        stm             r0,  {r4-r5}
  10.173 +        add             r0,  r0,  r2
  10.174 +        bne             1b
  10.175 +        pop             {r4-r5,pc}
  10.176 +        .align 5
  10.177 +2:
  10.178 +        ldm             r1,  {r4-r5, r12}
  10.179 +        add             r1,  r1,  r2
  10.180 +        ALIGN_DWORD     1,   r4,  r5,  r12
  10.181 +        pld             [r1]
  10.182 +        subs            r3,  r3,  #1
  10.183 +        stm             r0,  {r4-r5}
  10.184 +        add             r0,  r0,  r2
  10.185 +        bne             2b
  10.186 +        pop             {r4-r5,pc}
  10.187 +        .align 5
  10.188 +3:
  10.189 +        ldm             r1,  {r4-r5, r12}
  10.190 +        add             r1,  r1,  r2
  10.191 +        ALIGN_DWORD     2,   r4,  r5,  r12
  10.192 +        pld             [r1]
  10.193 +        subs            r3,  r3,  #1
  10.194 +        stm             r0,  {r4-r5}
  10.195 +        add             r0,  r0,  r2
  10.196 +        bne             3b
  10.197 +        pop             {r4-r5,pc}
  10.198 +        .align 5
  10.199 +4:
  10.200 +        ldm             r1,  {r4-r5, r12}
  10.201 +        add             r1,  r1,  r2
  10.202 +        ALIGN_DWORD     3,   r4,  r5,  r12
  10.203 +        pld             [r1]
  10.204 +        subs            r3,  r3,  #1
  10.205 +        stm             r0,  {r4-r5}
  10.206 +        add             r0,  r0,  r2
  10.207 +        bne             4b
  10.208 +        pop             {r4-r5,pc}
  10.209 +endfunc
  10.210 +
  10.211 +@ ----------------------------------------------------------------
  10.212 +        .align 5
  10.213 +function ff_put_pixels8_x2_arm, export=1
  10.214 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.215 +        @ block = word aligned, pixles = unaligned
  10.216 +        pld             [r1]
  10.217 +        push            {r4-r10,lr}
  10.218 +        ldr             r12, =0xfefefefe
  10.219 +        JMP_ALIGN       r5,  r1
  10.220 +1:
  10.221 +        ldm             r1,  {r4-r5, r10}
  10.222 +        add             r1,  r1,  r2
  10.223 +        ALIGN_DWORD_D   1,   r6,  r7,  r4,  r5,  r10
  10.224 +        pld             [r1]
  10.225 +        RND_AVG32       r8,  r9,  r4,  r5,  r6,  r7,  r12
  10.226 +        subs            r3,  r3,  #1
  10.227 +        stm             r0,  {r8-r9}
  10.228 +        add             r0,  r0,  r2
  10.229 +        bne             1b
  10.230 +        pop             {r4-r10,pc}
  10.231 +        .align 5
  10.232 +2:
  10.233 +        ldm             r1,  {r4-r5, r10}
  10.234 +        add             r1,  r1,  r2
  10.235 +        ALIGN_DWORD_D   1,   r6,  r7,  r4,  r5,  r10
  10.236 +        ALIGN_DWORD_D   2,   r8,  r9,  r4,  r5,  r10
  10.237 +        pld             [r1]
  10.238 +        RND_AVG32       r4,  r5,  r6,  r7,  r8,  r9,  r12
  10.239 +        subs            r3,  r3,  #1
  10.240 +        stm             r0,  {r4-r5}
  10.241 +        add             r0,  r0,  r2
  10.242 +        bne             2b
  10.243 +        pop             {r4-r10,pc}
  10.244 +        .align 5
  10.245 +3:
  10.246 +        ldm             r1,  {r4-r5, r10}
  10.247 +        add             r1,  r1,  r2
  10.248 +        ALIGN_DWORD_D   2,   r6,  r7,  r4,  r5,  r10
  10.249 +        ALIGN_DWORD_D   3,   r8,  r9,  r4,  r5,  r10
  10.250 +        pld             [r1]
  10.251 +        RND_AVG32       r4,  r5,  r6,  r7,  r8,  r9,  r12
  10.252 +        subs            r3,  r3,  #1
  10.253 +        stm             r0,  {r4-r5}
  10.254 +        add             r0,  r0,  r2
  10.255 +        bne             3b
  10.256 +        pop             {r4-r10,pc}
  10.257 +        .align 5
  10.258 +4:
  10.259 +        ldm             r1,  {r4-r5, r10}
  10.260 +        add             r1,  r1,  r2
  10.261 +        ALIGN_DWORD_D   3,   r6,  r7,  r4,  r5,  r10
  10.262 +        pld             [r1]
  10.263 +        RND_AVG32       r8,  r9,  r6,  r7,  r5,  r10, r12
  10.264 +        subs            r3,  r3,  #1
  10.265 +        stm             r0,  {r8-r9}
  10.266 +        add             r0,  r0,  r2
  10.267 +        bne             4b
  10.268 +        pop             {r4-r10,pc}
  10.269 +endfunc
  10.270 +
  10.271 +        .align 5
  10.272 +function ff_put_no_rnd_pixels8_x2_arm, export=1
  10.273 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.274 +        @ block = word aligned, pixles = unaligned
  10.275 +        pld             [r1]
  10.276 +        push            {r4-r10,lr}
  10.277 +        ldr             r12, =0xfefefefe
  10.278 +        JMP_ALIGN       r5,  r1
  10.279 +1:
  10.280 +        ldm             r1,  {r4-r5, r10}
  10.281 +        add             r1,  r1,  r2
  10.282 +        ALIGN_DWORD_D   1,   r6,  r7,  r4,  r5,  r10
  10.283 +        pld             [r1]
  10.284 +        NO_RND_AVG32    r8,  r9,  r4,  r5,  r6,  r7,  r12
  10.285 +        subs            r3,  r3,  #1
  10.286 +        stm             r0,  {r8-r9}
  10.287 +        add             r0,  r0,  r2
  10.288 +        bne             1b
  10.289 +        pop             {r4-r10,pc}
  10.290 +        .align 5
  10.291 +2:
  10.292 +        ldm             r1,  {r4-r5, r10}
  10.293 +        add             r1,  r1,  r2
  10.294 +        ALIGN_DWORD_D   1,   r6,  r7,  r4,  r5,  r10
  10.295 +        ALIGN_DWORD_D   2,   r8,  r9,  r4,  r5,  r10
  10.296 +        pld             [r1]
  10.297 +        NO_RND_AVG32    r4,  r5,  r6,  r7,  r8,  r9,  r12
  10.298 +        subs            r3,  r3,  #1
  10.299 +        stm             r0,  {r4-r5}
  10.300 +        add             r0,  r0,  r2
  10.301 +        bne             2b
  10.302 +        pop             {r4-r10,pc}
  10.303 +        .align 5
  10.304 +3:
  10.305 +        ldm             r1,  {r4-r5, r10}
  10.306 +        add             r1,  r1,  r2
  10.307 +        ALIGN_DWORD_D   2,   r6,  r7,  r4,  r5,  r10
  10.308 +        ALIGN_DWORD_D   3,   r8,  r9,  r4,  r5,  r10
  10.309 +        pld             [r1]
  10.310 +        NO_RND_AVG32    r4,  r5,  r6,  r7,  r8,  r9,  r12
  10.311 +        subs            r3,  r3,  #1
  10.312 +        stm             r0,  {r4-r5}
  10.313 +        add             r0,  r0,  r2
  10.314 +        bne             3b
  10.315 +        pop             {r4-r10,pc}
  10.316 +        .align 5
  10.317 +4:
  10.318 +        ldm             r1,  {r4-r5, r10}
  10.319 +        add             r1,  r1,  r2
  10.320 +        ALIGN_DWORD_D   3,   r6,  r7,  r4,  r5,  r10
  10.321 +        pld             [r1]
  10.322 +        NO_RND_AVG32    r8,  r9,  r6,  r7,  r5,  r10, r12
  10.323 +        subs            r3,  r3,  #1
  10.324 +        stm             r0,  {r8-r9}
  10.325 +        add             r0,  r0,  r2
  10.326 +        bne             4b
  10.327 +        pop             {r4-r10,pc}
  10.328 +endfunc
  10.329 +
  10.330 +
  10.331 +@ ----------------------------------------------------------------
  10.332 +        .align 5
  10.333 +function ff_put_pixels8_y2_arm, export=1
  10.334 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.335 +        @ block = word aligned, pixles = unaligned
  10.336 +        pld             [r1]
  10.337 +        push            {r4-r11,lr}
  10.338 +        mov             r3,  r3,  lsr #1
  10.339 +        ldr             r12, =0xfefefefe
  10.340 +        JMP_ALIGN       r5,  r1
  10.341 +1:
  10.342 +        ldm             r1,  {r4-r5}
  10.343 +        add             r1,  r1,  r2
  10.344 +6:      ldm             r1,  {r6-r7}
  10.345 +        add             r1,  r1,  r2
  10.346 +        pld             [r1]
  10.347 +        RND_AVG32       r8,  r9,  r4,  r5,  r6,  r7,  r12
  10.348 +        ldm             r1,  {r4-r5}
  10.349 +        add             r1,  r1,  r2
  10.350 +        stm             r0,  {r8-r9}
  10.351 +        add             r0,  r0,  r2
  10.352 +        pld             [r1]
  10.353 +        RND_AVG32       r8,  r9,  r6,  r7,  r4,  r5,  r12
  10.354 +        subs            r3,  r3,  #1
  10.355 +        stm             r0,  {r8-r9}
  10.356 +        add             r0,  r0,  r2
  10.357 +        bne             6b
  10.358 +        pop             {r4-r11,pc}
  10.359 +        .align 5
  10.360 +2:
  10.361 +        ldm             r1,  {r4-r6}
  10.362 +        add             r1,  r1,  r2
  10.363 +        pld             [r1]
  10.364 +        ALIGN_DWORD     1,   r4,  r5,  r6
  10.365 +6:      ldm             r1,  {r7-r9}
  10.366 +        add             r1,  r1,  r2
  10.367 +        pld             [r1]
  10.368 +        ALIGN_DWORD     1,   r7,  r8,  r9
  10.369 +        RND_AVG32       r10, r11, r4,  r5,  r7,  r8,  r12
  10.370 +        stm             r0,  {r10-r11}
  10.371 +        add             r0,  r0,  r2
  10.372 +        ldm             r1,  {r4-r6}
  10.373 +        add             r1,  r1,  r2
  10.374 +        pld             [r1]
  10.375 +        ALIGN_DWORD     1,   r4,  r5,  r6
  10.376 +        subs            r3,  r3,  #1
  10.377 +        RND_AVG32       r10, r11, r7,  r8,  r4,  r5,  r12
  10.378 +        stm             r0,  {r10-r11}
  10.379 +        add             r0,  r0,  r2
  10.380 +        bne             6b
  10.381 +        pop             {r4-r11,pc}
  10.382 +        .align 5
  10.383 +3:
  10.384 +        ldm             r1,  {r4-r6}
  10.385 +        add             r1,  r1,  r2
  10.386 +        pld             [r1]
  10.387 +        ALIGN_DWORD     2,   r4,  r5,  r6
  10.388 +6:      ldm             r1,  {r7-r9}
  10.389 +        add             r1,  r1,  r2
  10.390 +        pld             [r1]
  10.391 +        ALIGN_DWORD     2,   r7,  r8,  r9
  10.392 +        RND_AVG32       r10, r11, r4,  r5,  r7,  r8,  r12
  10.393 +        stm             r0,  {r10-r11}
  10.394 +        add             r0,  r0,  r2
  10.395 +        ldm             r1,  {r4-r6}
  10.396 +        add             r1,  r1,  r2
  10.397 +        pld             [r1]
  10.398 +        ALIGN_DWORD     2,   r4,  r5,  r6
  10.399 +        subs            r3,  r3,  #1
  10.400 +        RND_AVG32       r10, r11, r7,  r8,  r4,  r5,  r12
  10.401 +        stm             r0,  {r10-r11}
  10.402 +        add             r0,  r0,  r2
  10.403 +        bne             6b
  10.404 +        pop             {r4-r11,pc}
  10.405 +        .align 5
  10.406 +4:
  10.407 +        ldm             r1,  {r4-r6}
  10.408 +        add             r1,  r1,  r2
  10.409 +        pld             [r1]
  10.410 +        ALIGN_DWORD     3,   r4,  r5,  r6
  10.411 +6:      ldm             r1,  {r7-r9}
  10.412 +        add             r1,  r1,  r2
  10.413 +        pld             [r1]
  10.414 +        ALIGN_DWORD     3,   r7,  r8,  r9
  10.415 +        RND_AVG32       r10, r11, r4,  r5,  r7,  r8,  r12
  10.416 +        stm             r0,  {r10-r11}
  10.417 +        add             r0,  r0,  r2
  10.418 +        ldm             r1,  {r4-r6}
  10.419 +        add             r1,  r1,  r2
  10.420 +        pld             [r1]
  10.421 +        ALIGN_DWORD     3,   r4,  r5,  r6
  10.422 +        subs            r3,  r3,  #1
  10.423 +        RND_AVG32       r10, r11, r7,  r8,  r4,  r5,  r12
  10.424 +        stm             r0,  {r10-r11}
  10.425 +        add             r0,  r0,  r2
  10.426 +        bne             6b
  10.427 +        pop             {r4-r11,pc}
  10.428 +endfunc
  10.429 +
  10.430 +        .align 5
  10.431 +function ff_put_no_rnd_pixels8_y2_arm, export=1
  10.432 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.433 +        @ block = word aligned, pixles = unaligned
  10.434 +        pld             [r1]
  10.435 +        push            {r4-r11,lr}
  10.436 +        mov             r3,  r3,  lsr #1
  10.437 +        ldr             r12, =0xfefefefe
  10.438 +        JMP_ALIGN       r5,  r1
  10.439 +1:
  10.440 +        ldm             r1,  {r4-r5}
  10.441 +        add             r1,  r1,  r2
  10.442 +6:      ldm             r1,  {r6-r7}
  10.443 +        add             r1,  r1,  r2
  10.444 +        pld             [r1]
  10.445 +        NO_RND_AVG32    r8,  r9,  r4,  r5,  r6,  r7,  r12
  10.446 +        ldm             r1,  {r4-r5}
  10.447 +        add             r1,  r1,  r2
  10.448 +        stm             r0,  {r8-r9}
  10.449 +        add             r0,  r0,  r2
  10.450 +        pld             [r1]
  10.451 +        NO_RND_AVG32    r8,  r9,  r6,  r7,  r4,  r5,  r12
  10.452 +        subs            r3,  r3,  #1
  10.453 +        stm             r0,  {r8-r9}
  10.454 +        add             r0,  r0,  r2
  10.455 +        bne             6b
  10.456 +        pop             {r4-r11,pc}
  10.457 +        .align 5
  10.458 +2:
  10.459 +        ldm             r1,  {r4-r6}
  10.460 +        add             r1,  r1,  r2
  10.461 +        pld             [r1]
  10.462 +        ALIGN_DWORD     1,   r4,  r5,  r6
  10.463 +6:      ldm             r1,  {r7-r9}
  10.464 +        add             r1,  r1,  r2
  10.465 +        pld             [r1]
  10.466 +        ALIGN_DWORD     1,   r7,  r8,  r9
  10.467 +        NO_RND_AVG32    r10, r11, r4,  r5,  r7,  r8,  r12
  10.468 +        stm             r0,  {r10-r11}
  10.469 +        add             r0,  r0,  r2
  10.470 +        ldm             r1,  {r4-r6}
  10.471 +        add             r1,  r1,  r2
  10.472 +        pld             [r1]
  10.473 +        ALIGN_DWORD     1,   r4,  r5,  r6
  10.474 +        subs            r3,  r3,  #1
  10.475 +        NO_RND_AVG32    r10, r11, r7,  r8,  r4,  r5,  r12
  10.476 +        stm             r0,  {r10-r11}
  10.477 +        add             r0,  r0,  r2
  10.478 +        bne             6b
  10.479 +        pop             {r4-r11,pc}
  10.480 +        .align 5
  10.481 +3:
  10.482 +        ldm             r1,  {r4-r6}
  10.483 +        add             r1,  r1,  r2
  10.484 +        pld             [r1]
  10.485 +        ALIGN_DWORD     2,   r4,  r5,  r6
  10.486 +6:      ldm             r1,  {r7-r9}
  10.487 +        add             r1,  r1,  r2
  10.488 +        pld             [r1]
  10.489 +        ALIGN_DWORD     2,   r7,  r8,  r9
  10.490 +        NO_RND_AVG32    r10, r11, r4,  r5,  r7,  r8,  r12
  10.491 +        stm             r0,  {r10-r11}
  10.492 +        add             r0,  r0,  r2
  10.493 +        ldm             r1,  {r4-r6}
  10.494 +        add             r1,  r1,  r2
  10.495 +        pld             [r1]
  10.496 +        ALIGN_DWORD     2,   r4,  r5,  r6
  10.497 +        subs            r3,  r3,  #1
  10.498 +        NO_RND_AVG32    r10, r11, r7,  r8,  r4,  r5,  r12
  10.499 +        stm             r0,  {r10-r11}
  10.500 +        add             r0,  r0,  r2
  10.501 +        bne             6b
  10.502 +        pop             {r4-r11,pc}
  10.503 +        .align 5
  10.504 +4:
  10.505 +        ldm             r1,  {r4-r6}
  10.506 +        add             r1,  r1,  r2
  10.507 +        pld             [r1]
  10.508 +        ALIGN_DWORD     3,   r4,  r5,  r6
  10.509 +6:      ldm             r1,  {r7-r9}
  10.510 +        add             r1,  r1,  r2
  10.511 +        pld             [r1]
  10.512 +        ALIGN_DWORD     3,   r7,  r8,  r9
  10.513 +        NO_RND_AVG32    r10, r11, r4,  r5,  r7,  r8,  r12
  10.514 +        stm             r0,  {r10-r11}
  10.515 +        add             r0,  r0,  r2
  10.516 +        ldm             r1,  {r4-r6}
  10.517 +        add             r1,  r1,  r2
  10.518 +        pld             [r1]
  10.519 +        ALIGN_DWORD     3,   r4,  r5,  r6
  10.520 +        subs            r3,  r3,  #1
  10.521 +        NO_RND_AVG32    r10, r11, r7,  r8,  r4,  r5,  r12
  10.522 +        stm             r0,  {r10-r11}
  10.523 +        add             r0,  r0,  r2
  10.524 +        bne             6b
  10.525 +        pop             {r4-r11,pc}
  10.526 +endfunc
  10.527 +
  10.528 +        .ltorg
  10.529 +
  10.530 +@ ----------------------------------------------------------------
  10.531 +.macro  RND_XY2_IT align, rnd
  10.532 +        @ l1=  (a & 0x03030303) + (b & 0x03030303) ?(+ 0x02020202)
  10.533 +        @ h1= ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2)
  10.534 +.if \align == 0
  10.535 +        ldm             r1,  {r6-r8}
  10.536 +.elseif \align == 3
  10.537 +        ldm             r1,  {r5-r7}
  10.538 +.else
  10.539 +        ldm             r1,  {r8-r10}
  10.540 +.endif
  10.541 +        add             r1,  r1,  r2
  10.542 +        pld             [r1]
  10.543 +.if \align == 0
  10.544 +        ALIGN_DWORD_D   1,   r4,  r5,  r6,  r7,  r8
  10.545 +.elseif \align == 1
  10.546 +        ALIGN_DWORD_D   1,   r4,  r5,  r8,  r9,  r10
  10.547 +        ALIGN_DWORD_D   2,   r6,  r7,  r8,  r9,  r10
  10.548 +.elseif \align == 2
  10.549 +        ALIGN_DWORD_D   2,   r4,  r5,  r8,  r9,  r10
  10.550 +        ALIGN_DWORD_D   3,   r6,  r7,  r8,  r9,  r10
  10.551 +.elseif \align == 3
  10.552 +        ALIGN_DWORD_D   3,   r4,  r5,  r5,  r6,  r7
  10.553 +.endif
  10.554 +        ldr             r14, =0x03030303
  10.555 +        tst             r3,  #1
  10.556 +        and             r8,  r4,  r14
  10.557 +        and             r9,  r5,  r14
  10.558 +        and             r10, r6,  r14
  10.559 +        and             r11, r7,  r14
  10.560 +        andeq           r14, r14, r14, \rnd #1
  10.561 +        add             r8,  r8,  r10
  10.562 +        add             r9,  r9,  r11
  10.563 +        ldr             r12, =0xfcfcfcfc >> 2
  10.564 +        addeq           r8,  r8,  r14
  10.565 +        addeq           r9,  r9,  r14
  10.566 +        and             r4,  r12, r4,  lsr #2
  10.567 +        and             r5,  r12, r5,  lsr #2
  10.568 +        and             r6,  r12, r6,  lsr #2
  10.569 +        and             r7,  r12, r7,  lsr #2
  10.570 +        add             r10, r4,  r6
  10.571 +        add             r11, r5,  r7
  10.572 +        subs            r3,  r3,  #1
  10.573 +.endm
  10.574 +
  10.575 +.macro RND_XY2_EXPAND align, rnd
  10.576 +        RND_XY2_IT      \align, \rnd
  10.577 +6:      push            {r8-r11}
  10.578 +        RND_XY2_IT      \align, \rnd
  10.579 +        pop             {r4-r7}
  10.580 +        add             r4,  r4,  r8
  10.581 +        add             r5,  r5,  r9
  10.582 +        ldr             r14, =0x0f0f0f0f
  10.583 +        add             r6,  r6,  r10
  10.584 +        add             r7,  r7,  r11
  10.585 +        and             r4,  r14, r4,  lsr #2
  10.586 +        and             r5,  r14, r5,  lsr #2
  10.587 +        add             r4,  r4,  r6
  10.588 +        add             r5,  r5,  r7
  10.589 +        stm             r0,  {r4-r5}
  10.590 +        add             r0,  r0,  r2
  10.591 +        bge             6b
  10.592 +        pop             {r4-r11,pc}
  10.593 +.endm
  10.594 +
  10.595 +        .align 5
  10.596 +function ff_put_pixels8_xy2_arm, export=1
  10.597 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.598 +        @ block = word aligned, pixles = unaligned
  10.599 +        pld             [r1]
  10.600 +        push            {r4-r11,lr} @ R14 is also called LR
  10.601 +        JMP_ALIGN       r5,  r1
  10.602 +1:      RND_XY2_EXPAND  0, lsl
  10.603 +        .align 5
  10.604 +2:      RND_XY2_EXPAND  1, lsl
  10.605 +        .align 5
  10.606 +3:      RND_XY2_EXPAND  2, lsl
  10.607 +        .align 5
  10.608 +4:      RND_XY2_EXPAND  3, lsl
  10.609 +endfunc
  10.610 +
  10.611 +        .align 5
  10.612 +function ff_put_no_rnd_pixels8_xy2_arm, export=1
  10.613 +        @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h)
  10.614 +        @ block = word aligned, pixles = unaligned
  10.615 +        pld             [r1]
  10.616 +        push            {r4-r11,lr}
  10.617 +        JMP_ALIGN       r5,  r1
  10.618 +1:      RND_XY2_EXPAND  0, lsr
  10.619 +        .align 5
  10.620 +2:      RND_XY2_EXPAND  1, lsr
  10.621 +        .align 5
  10.622 +3:      RND_XY2_EXPAND  2, lsr
  10.623 +        .align 5
  10.624 +4:      RND_XY2_EXPAND  3, lsr
  10.625 +endfunc
  10.626 +
  10.627 +        .align 5
  10.628 +@ void ff_add_pixels_clamped_arm(int16_t *block, uint8_t *dest, int stride)
  10.629 +function ff_add_pixels_clamped_arm, export=1
  10.630 +        push            {r4-r10}
  10.631 +        mov             r10, #8
  10.632 +1:
  10.633 +        ldr             r4,  [r1]               /* load dest */
  10.634 +        /* block[0] and block[1]*/
  10.635 +        ldrsh           r5,  [r0]
  10.636 +        ldrsh           r7,  [r0, #2]
  10.637 +        and             r6,  r4,  #0xFF
  10.638 +        and             r8,  r4,  #0xFF00
  10.639 +        add             r6,  r5,  r6
  10.640 +        add             r8,  r7,  r8,  lsr #8
  10.641 +        mvn             r5,  r5
  10.642 +        mvn             r7,  r7
  10.643 +        tst             r6,  #0x100
  10.644 +        movne           r6,  r5,  lsr #24
  10.645 +        tst             r8,  #0x100
  10.646 +        movne           r8,  r7,  lsr #24
  10.647 +        mov             r9,  r6
  10.648 +        ldrsh           r5,  [r0, #4]           /* moved form [A] */
  10.649 +        orr             r9,  r9,  r8,  lsl #8
  10.650 +        /* block[2] and block[3] */
  10.651 +        /* [A] */
  10.652 +        ldrsh           r7,  [r0, #6]
  10.653 +        and             r6,  r4,  #0xFF0000
  10.654 +        and             r8,  r4,  #0xFF000000
  10.655 +        add             r6,  r5,  r6,  lsr #16
  10.656 +        add             r8,  r7,  r8,  lsr #24
  10.657 +        mvn             r5,  r5
  10.658 +        mvn             r7,  r7
  10.659 +        tst             r6,  #0x100
  10.660 +        movne           r6,  r5,  lsr #24
  10.661 +        tst             r8,  #0x100
  10.662 +        movne           r8,  r7,  lsr #24
  10.663 +        orr             r9,  r9,  r6,  lsl #16
  10.664 +        ldr             r4,  [r1, #4]           /* moved form [B] */
  10.665 +        orr             r9,  r9,  r8,  lsl #24
  10.666 +        /* store dest */
  10.667 +        ldrsh           r5,  [r0, #8]           /* moved form [C] */
  10.668 +        str             r9,  [r1]
  10.669 +
  10.670 +        /* load dest */
  10.671 +        /* [B] */
  10.672 +        /* block[4] and block[5] */
  10.673 +        /* [C] */
  10.674 +        ldrsh           r7,  [r0, #10]
  10.675 +        and             r6,  r4,  #0xFF
  10.676 +        and             r8,  r4,  #0xFF00
  10.677 +        add             r6,  r5,  r6
  10.678 +        add             r8,  r7,  r8,  lsr #8
  10.679 +        mvn             r5,  r5
  10.680 +        mvn             r7,  r7
  10.681 +        tst             r6,  #0x100
  10.682 +        movne           r6,  r5,  lsr #24
  10.683 +        tst             r8,  #0x100
  10.684 +        movne           r8,  r7,  lsr #24
  10.685 +        mov             r9,  r6
  10.686 +        ldrsh           r5,  [r0, #12]          /* moved from [D] */
  10.687 +        orr             r9,  r9,  r8,  lsl #8
  10.688 +        /* block[6] and block[7] */
  10.689 +        /* [D] */
  10.690 +        ldrsh           r7,  [r0, #14]
  10.691 +        and             r6,  r4,  #0xFF0000
  10.692 +        and             r8,  r4,  #0xFF000000
  10.693 +        add             r6,  r5,  r6,  lsr #16
  10.694 +        add             r8,  r7,  r8,  lsr #24
  10.695 +        mvn             r5,  r5
  10.696 +        mvn             r7,  r7
  10.697 +        tst             r6,  #0x100
  10.698 +        movne           r6,  r5,  lsr #24
  10.699 +        tst             r8,  #0x100
  10.700 +        movne           r8,  r7,  lsr #24
  10.701 +        orr             r9,  r9,  r6,  lsl #16
  10.702 +        add             r0,  r0,  #16           /* moved from [E] */
  10.703 +        orr             r9,  r9,  r8,  lsl #24
  10.704 +        subs            r10, r10, #1            /* moved from [F] */
  10.705 +        /* store dest */
  10.706 +        str             r9,  [r1, #4]
  10.707 +
  10.708 +        /* [E] */
  10.709 +        /* [F] */
  10.710 +        add             r1,  r1,  r2
  10.711 +        bne             1b
  10.712 +
  10.713 +        pop             {r4-r10}
  10.714 +        bx              lr
  10.715 +endfunc

    11.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
    11.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dsputil_arm.h	Mon Aug 27 12:09:56 2012 +0200
    11.3 @@ -0,0 +1,33 @@
    11.4 +/*
    11.5 + * Copyright (c) 2009 Mans Rullgard <mans@mansr.com>
    11.6 + *
    11.7 + * This file is part of FFmpeg.
    11.8 + *
    11.9 + * FFmpeg is free software; you can redistribute it and/or
   11.10 + * modify it under the terms of the GNU Lesser General Public
   11.11 + * License as published by the Free Software Foundation; either
   11.12 + * version 2.1 of the License, or (at your option) any later version.
   11.13 + *
   11.14 + * FFmpeg is distributed in the hope that it will be useful,
   11.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
   11.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   11.17 + * Lesser General Public License for more details.
   11.18 + *
   11.19 + * You should have received a copy of the GNU Lesser General Public
   11.20 + * License along with FFmpeg; if not, write to the Free Software
   11.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
   11.22 + */
   11.23 +
   11.24 +#ifndef AVCODEC_ARM_DSPUTIL_H
   11.25 +#define AVCODEC_ARM_DSPUTIL_H
   11.26 +
   11.27 +#include "libavcodec/avcodec.h"
   11.28 +#include "libavcodec/dsputil.h"
   11.29 +
   11.30 +void ff_dsputil_init_armv5te(DSPContext* c, AVCodecContext *avctx);
   11.31 +void ff_dsputil_init_armv6(DSPContext* c, AVCodecContext *avctx);
   11.32 +void ff_dsputil_init_vfp(DSPContext* c, AVCodecContext *avctx);
   11.33 +void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx);
   11.34 +void ff_dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx);
   11.35 +
   11.36 +#endif

    12.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
    12.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dsputil_armv6.S	Mon Aug 27 12:09:56 2012 +0200
    12.3 @@ -0,0 +1,623 @@
    12.4 +/*
    12.5 + * Copyright (c) 2009 Mans Rullgard <mans@mansr.com>
    12.6 + *
    12.7 + * This file is part of FFmpeg.
    12.8 + *
    12.9 + * FFmpeg is free software; you can redistribute it and/or
   12.10 + * modify it under the terms of the GNU Lesser General Public
   12.11 + * License as published by the Free Software Foundation; either
   12.12 + * version 2.1 of the License, or (at your option) any later version.
   12.13 + *
   12.14 + * FFmpeg is distributed in the hope that it will be useful,
   12.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
   12.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   12.17 + * Lesser General Public License for more details.
   12.18 + *
   12.19 + * You should have received a copy of the GNU Lesser General Public
   12.20 + * License along with FFmpeg; if not, write to the Free Software
   12.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
   12.22 + */
   12.23 +
   12.24 +#include "asm.S"
   12.25 +
   12.26 +        preserve8
   12.27 +
   12.28 +        .text
   12.29 +
   12.30 +.macro  call_2x_pixels  type, subp
   12.31 +function ff_\type\()_pixels16\subp\()_armv6, export=1
   12.32 +        push            {r0-r3, lr}
   12.33 +        bl              ff_\type\()_pixels8\subp\()_armv6
   12.34 +        pop             {r0-r3, lr}
   12.35 +        add             r0,  r0,  #8
   12.36 +        add             r1,  r1,  #8
   12.37 +        b               ff_\type\()_pixels8\subp\()_armv6
   12.38 +endfunc
   12.39 +.endm
   12.40 +
   12.41 +call_2x_pixels          avg
   12.42 +call_2x_pixels          put, _x2
   12.43 +call_2x_pixels          put, _y2
   12.44 +call_2x_pixels          put, _x2_no_rnd
   12.45 +call_2x_pixels          put, _y2_no_rnd
   12.46 +
   12.47 +function ff_put_pixels16_armv6, export=1
   12.48 +        push            {r4-r11}
   12.49 +1:
   12.50 +        ldr             r5,  [r1, #4]
   12.51 +        ldr             r6,  [r1, #8]
   12.52 +        ldr             r7,  [r1, #12]
   12.53 +        ldr             r4,  [r1], r2
   12.54 +        strd            r6,  r7,  [r0, #8]
   12.55 +        ldr             r9,  [r1, #4]
   12.56 +        strd            r4,  r5,  [r0],  r2
   12.57 +        ldr             r10, [r1, #8]
   12.58 +        ldr             r11, [r1, #12]
   12.59 +        ldr             r8,  [r1], r2
   12.60 +        strd            r10, r11, [r0, #8]
   12.61 +        subs            r3,  r3,  #2
   12.62 +        strd            r8,  r9,  [r0],  r2
   12.63 +        bne             1b
   12.64 +
   12.65 +        pop             {r4-r11}
   12.66 +        bx              lr
   12.67 +endfunc
   12.68 +
   12.69 +function ff_put_pixels8_armv6, export=1
   12.70 +        push            {r4-r7}
   12.71 +1:
   12.72 +        ldr             r5,  [r1, #4]
   12.73 +        ldr             r4,  [r1], r2
   12.74 +        ldr             r7,  [r1, #4]
   12.75 +        strd            r4,  r5,  [r0],  r2
   12.76 +        ldr             r6,  [r1], r2
   12.77 +        subs            r3,  r3,  #2
   12.78 +        strd            r6,  r7,  [r0],  r2
   12.79 +        bne             1b
   12.80 +
   12.81 +        pop             {r4-r7}
   12.82 +        bx              lr
   12.83 +endfunc
   12.84 +
   12.85 +function ff_put_pixels8_x2_armv6, export=1
   12.86 +        push            {r4-r11, lr}
   12.87 +        mov             r12, #1
   12.88 +        orr             r12, r12, r12, lsl #8
   12.89 +        orr             r12, r12, r12, lsl #16
   12.90 +1:
   12.91 +        ldr             r4,  [r1]
   12.92 +        subs            r3,  r3,  #2
   12.93 +        ldr             r5,  [r1, #4]
   12.94 +        ldr             r7,  [r1, #5]
   12.95 +        lsr             r6,  r4,  #8
   12.96 +        ldr             r8,  [r1, r2]!
   12.97 +        orr             r6,  r6,  r5,  lsl #24
   12.98 +        ldr             r9,  [r1, #4]
   12.99 +        ldr             r11, [r1, #5]
  12.100 +        lsr             r10, r8,  #8
  12.101 +        add             r1,  r1,  r2
  12.102 +        orr             r10, r10, r9,  lsl #24
  12.103 +        eor             r14, r4,  r6
  12.104 +        uhadd8          r4,  r4,  r6
  12.105 +        eor             r6,  r5,  r7
  12.106 +        uhadd8          r5,  r5,  r7
  12.107 +        and             r14, r14, r12
  12.108 +        and             r6,  r6,  r12
  12.109 +        uadd8           r4,  r4,  r14
  12.110 +        eor             r14, r8,  r10
  12.111 +        uadd8           r5,  r5,  r6
  12.112 +        eor             r6,  r9,  r11
  12.113 +        uhadd8          r8,  r8,  r10
  12.114 +        and             r14, r14, r12
  12.115 +        uhadd8          r9,  r9,  r11
  12.116 +        and             r6,  r6,  r12
  12.117 +        uadd8           r8,  r8,  r14
  12.118 +        strd            r4,  r5,  [r0],  r2
  12.119 +        uadd8           r9,  r9,  r6
  12.120 +        strd            r8,  r9,  [r0],  r2
  12.121 +        bne             1b
  12.122 +
  12.123 +        pop             {r4-r11, pc}
  12.124 +endfunc
  12.125 +
  12.126 +function ff_put_pixels8_y2_armv6, export=1
  12.127 +        push            {r4-r11}
  12.128 +        mov             r12, #1
  12.129 +        orr             r12, r12, r12, lsl #8
  12.130 +        orr             r12, r12, r12, lsl #16
  12.131 +        ldr             r4,  [r1]
  12.132 +        ldr             r5,  [r1, #4]
  12.133 +        ldr             r6,  [r1, r2]!
  12.134 +        ldr             r7,  [r1, #4]
  12.135 +1:
  12.136 +        subs            r3,  r3,  #2
  12.137 +        uhadd8          r8,  r4,  r6
  12.138 +        eor             r10, r4,  r6
  12.139 +        uhadd8          r9,  r5,  r7
  12.140 +        eor             r11, r5,  r7
  12.141 +        and             r10, r10, r12
  12.142 +        ldr             r4,  [r1, r2]!
  12.143 +        uadd8           r8,  r8,  r10
  12.144 +        and             r11, r11, r12
  12.145 +        uadd8           r9,  r9,  r11
  12.146 +        ldr             r5,  [r1, #4]
  12.147 +        uhadd8          r10, r4,  r6
  12.148 +        eor             r6,  r4,  r6
  12.149 +        uhadd8          r11, r5,  r7
  12.150 +        and             r6,  r6,  r12
  12.151 +        eor             r7,  r5,  r7
  12.152 +        uadd8           r10, r10, r6
  12.153 +        and             r7,  r7,  r12
  12.154 +        ldr             r6,  [r1, r2]!
  12.155 +        uadd8           r11, r11, r7
  12.156 +        strd            r8,  r9,  [r0],  r2
  12.157 +        ldr             r7,  [r1, #4]
  12.158 +        strd            r10, r11, [r0],  r2
  12.159 +        bne             1b
  12.160 +
  12.161 +        pop             {r4-r11}
  12.162 +        bx              lr
  12.163 +endfunc
  12.164 +
  12.165 +function ff_put_pixels8_x2_no_rnd_armv6, export=1
  12.166 +        push            {r4-r9, lr}
  12.167 +1:
  12.168 +        subs            r3,  r3,  #2
  12.169 +        ldr             r4,  [r1]
  12.170 +        ldr             r5,  [r1, #4]
  12.171 +        ldr             r7,  [r1, #5]
  12.172 +        ldr             r8,  [r1, r2]!
  12.173 +        ldr             r9,  [r1, #4]
  12.174 +        ldr             r14, [r1, #5]
  12.175 +        add             r1,  r1,  r2
  12.176 +        lsr             r6,  r4,  #8
  12.177 +        orr             r6,  r6,  r5,  lsl #24
  12.178 +        lsr             r12, r8,  #8
  12.179 +        orr             r12, r12, r9,  lsl #24
  12.180 +        uhadd8          r4,  r4,  r6
  12.181 +        uhadd8          r5,  r5,  r7
  12.182 +        uhadd8          r8,  r8,  r12
  12.183 +        uhadd8          r9,  r9,  r14
  12.184 +        stm             r0,  {r4,r5}
  12.185 +        add             r0,  r0,  r2
  12.186 +        stm             r0,  {r8,r9}
  12.187 +        add             r0,  r0,  r2
  12.188 +        bne             1b
  12.189 +
  12.190 +        pop             {r4-r9, pc}
  12.191 +endfunc
  12.192 +
  12.193 +function ff_put_pixels8_y2_no_rnd_armv6, export=1
  12.194 +        push            {r4-r9, lr}
  12.195 +        ldr             r4,  [r1]
  12.196 +        ldr             r5,  [r1, #4]
  12.197 +        ldr             r6,  [r1, r2]!
  12.198 +        ldr             r7,  [r1, #4]
  12.199 +1:
  12.200 +        subs            r3,  r3,  #2
  12.201 +        uhadd8          r8,  r4,  r6
  12.202 +        ldr             r4,  [r1, r2]!
  12.203 +        uhadd8          r9,  r5,  r7
  12.204 +        ldr             r5,  [r1, #4]
  12.205 +        uhadd8          r12, r4,  r6
  12.206 +        ldr             r6,  [r1, r2]!
  12.207 +        uhadd8          r14, r5,  r7
  12.208 +        ldr             r7,  [r1, #4]
  12.209 +        stm             r0,  {r8,r9}
  12.210 +        add             r0,  r0,  r2
  12.211 +        stm             r0,  {r12,r14}
  12.212 +        add             r0,  r0,  r2
  12.213 +        bne             1b
  12.214 +
  12.215 +        pop             {r4-r9, pc}
  12.216 +endfunc
  12.217 +
  12.218 +function ff_avg_pixels8_armv6, export=1
  12.219 +        pld             [r1, r2]
  12.220 +        push            {r4-r10, lr}
  12.221 +        mov             lr,  #1
  12.222 +        orr             lr,  lr,  lr,  lsl #8
  12.223 +        orr             lr,  lr,  lr,  lsl #16
  12.224 +        ldrd            r4,  r5,  [r0]
  12.225 +        ldr             r10, [r1, #4]
  12.226 +        ldr             r9,  [r1], r2
  12.227 +        subs            r3,  r3,  #2
  12.228 +1:
  12.229 +        pld             [r1, r2]
  12.230 +        eor             r8,  r4,  r9
  12.231 +        uhadd8          r4,  r4,  r9
  12.232 +        eor             r12, r5,  r10
  12.233 +        ldrd            r6,  r7,  [r0, r2]
  12.234 +        uhadd8          r5,  r5,  r10
  12.235 +        and             r8,  r8,  lr
  12.236 +        ldr             r10, [r1, #4]
  12.237 +        and             r12, r12, lr
  12.238 +        uadd8           r4,  r4,  r8
  12.239 +        ldr             r9,  [r1], r2
  12.240 +        eor             r8,  r6,  r9
  12.241 +        uadd8           r5,  r5,  r12
  12.242 +        pld             [r1, r2,  lsl #1]
  12.243 +        eor             r12, r7,  r10
  12.244 +        uhadd8          r6,  r6,  r9
  12.245 +        strd            r4,  r5,  [r0], r2
  12.246 +        uhadd8          r7,  r7,  r10
  12.247 +        beq             2f
  12.248 +        and             r8,  r8,  lr
  12.249 +        ldrd            r4,  r5,  [r0, r2]
  12.250 +        uadd8           r6,  r6,  r8
  12.251 +        ldr             r10, [r1, #4]
  12.252 +        and             r12, r12, lr
  12.253 +        subs            r3,  r3,  #2
  12.254 +        uadd8           r7,  r7,  r12
  12.255 +        ldr             r9,  [r1], r2
  12.256 +        strd            r6,  r7,  [r0], r2
  12.257 +        b               1b
  12.258 +2:
  12.259 +        and             r8,  r8,  lr
  12.260 +        and             r12, r12, lr
  12.261 +        uadd8           r6,  r6,  r8
  12.262 +        uadd8           r7,  r7,  r12
  12.263 +        strd            r6,  r7,  [r0], r2
  12.264 +
  12.265 +        pop             {r4-r10, pc}
  12.266 +endfunc
  12.267 +
  12.268 +function ff_add_pixels_clamped_armv6, export=1
  12.269 +        push            {r4-r8,lr}
  12.270 +        mov             r3,  #8
  12.271 +1:
  12.272 +        ldm             r0!, {r4,r5,r12,lr}
  12.273 +        ldrd            r6,  r7,  [r1]
  12.274 +        pkhbt           r8,  r4,  r5,  lsl #16
  12.275 +        pkhtb           r5,  r5,  r4,  asr #16
  12.276 +        pkhbt           r4,  r12, lr,  lsl #16
  12.277 +        pkhtb           lr,  lr,  r12, asr #16
  12.278 +        pld             [r1, r2]
  12.279 +        uxtab16         r8,  r8,  r6
  12.280 +        uxtab16         r5,  r5,  r6,  ror #8
  12.281 +        uxtab16         r4,  r4,  r7
  12.282 +        uxtab16         lr,  lr,  r7,  ror #8
  12.283 +        usat16          r8,  #8,  r8
  12.284 +        usat16          r5,  #8,  r5
  12.285 +        usat16          r4,  #8,  r4
  12.286 +        usat16          lr,  #8,  lr
  12.287 +        orr             r6,  r8,  r5,  lsl #8
  12.288 +        orr             r7,  r4,  lr,  lsl #8
  12.289 +        subs            r3,  r3,  #1
  12.290 +        strd            r6,  r7,  [r1],  r2
  12.291 +        bgt             1b
  12.292 +        pop             {r4-r8,pc}
  12.293 +endfunc
  12.294 +
  12.295 +function ff_get_pixels_armv6, export=1
  12.296 +        pld             [r1, r2]
  12.297 +        push            {r4-r8, lr}
  12.298 +        mov             lr,  #8
  12.299 +1:
  12.300 +        ldrd            r4,  r5,  [r1],  r2
  12.301 +        subs            lr,  lr,  #1
  12.302 +        uxtb16          r6,  r4
  12.303 +        uxtb16          r4,  r4,  ror #8
  12.304 +        uxtb16          r12, r5
  12.305 +        uxtb16          r8,  r5,  ror #8
  12.306 +        pld             [r1, r2]
  12.307 +        pkhbt           r5,  r6,  r4,  lsl #16
  12.308 +        pkhtb           r6,  r4,  r6,  asr #16
  12.309 +        pkhbt           r7,  r12, r8,  lsl #16
  12.310 +        pkhtb           r12, r8,  r12, asr #16
  12.311 +        stm             r0!, {r5,r6,r7,r12}
  12.312 +        bgt             1b
  12.313 +
  12.314 +        pop             {r4-r8, pc}
  12.315 +endfunc
  12.316 +
  12.317 +function ff_diff_pixels_armv6, export=1
  12.318 +        pld             [r1, r3]
  12.319 +        pld             [r2, r3]
  12.320 +        push            {r4-r9, lr}
  12.321 +        mov             lr,  #8
  12.322 +1:
  12.323 +        ldrd            r4,  r5,  [r1],  r3
  12.324 +        ldrd            r6,  r7,  [r2],  r3
  12.325 +        uxtb16          r8,  r4
  12.326 +        uxtb16          r4,  r4,  ror #8
  12.327 +        uxtb16          r9,  r6
  12.328 +        uxtb16          r6,  r6,  ror #8
  12.329 +        pld             [r1, r3]
  12.330 +        ssub16          r9,  r8,  r9
  12.331 +        ssub16          r6,  r4,  r6
  12.332 +        uxtb16          r8,  r5
  12.333 +        uxtb16          r5,  r5,  ror #8
  12.334 +        pld             [r2, r3]
  12.335 +        pkhbt           r4,  r9,  r6,  lsl #16
  12.336 +        pkhtb           r6,  r6,  r9,  asr #16
  12.337 +        uxtb16          r9,  r7
  12.338 +        uxtb16          r7,  r7,  ror #8
  12.339 +        ssub16          r9,  r8,  r9
  12.340 +        ssub16          r5,  r5,  r7
  12.341 +        subs            lr,  lr,  #1
  12.342 +        pkhbt           r8,  r9,  r5,  lsl #16
  12.343 +        pkhtb           r9,  r5,  r9,  asr #16
  12.344 +        stm             r0!, {r4,r6,r8,r9}
  12.345 +        bgt             1b
  12.346 +
  12.347 +        pop             {r4-r9, pc}
  12.348 +endfunc
  12.349 +
  12.350 +function ff_pix_abs16_armv6, export=1
  12.351 +        ldr             r0,  [sp]
  12.352 +        push            {r4-r9, lr}
  12.353 +        mov             r12, #0
  12.354 +        mov             lr,  #0
  12.355 +        ldm             r1,  {r4-r7}
  12.356 +        ldr             r8,  [r2]
  12.357 +1:
  12.358 +        ldr             r9,  [r2, #4]
  12.359 +        pld             [r1, r3]
  12.360 +        usada8          r12, r4,  r8,  r12
  12.361 +        ldr             r8,  [r2, #8]
  12.362 +        pld             [r2, r3]
  12.363 +        usada8          lr,  r5,  r9,  lr
  12.364 +        ldr             r9,  [r2, #12]
  12.365 +        usada8          r12, r6,  r8,  r12
  12.366 +        subs            r0,  r0,  #1
  12.367 +        usada8          lr,  r7,  r9,  lr
  12.368 +        beq             2f
  12.369 +        add             r1,  r1,  r3
  12.370 +        ldm             r1,  {r4-r7}
  12.371 +        add             r2,  r2,  r3
  12.372 +        ldr             r8,  [r2]
  12.373 +        b               1b
  12.374 +2:
  12.375 +        add             r0,  r12, lr
  12.376 +        pop             {r4-r9, pc}
  12.377 +endfunc
  12.378 +
  12.379 +function ff_pix_abs16_x2_armv6, export=1
  12.380 +        ldr             r12, [sp]
  12.381 +        push            {r4-r11, lr}
  12.382 +        mov             r0,  #0
  12.383 +        mov             lr,  #1
  12.384 +        orr             lr,  lr,  lr,  lsl #8
  12.385 +        orr             lr,  lr,  lr,  lsl #16
  12.386 +1:
  12.387 +        ldr             r8,  [r2]
  12.388 +        ldr             r9,  [r2, #4]
  12.389 +        lsr             r10, r8,  #8
  12.390 +        ldr             r4,  [r1]
  12.391 +        lsr             r6,  r9,  #8
  12.392 +        orr             r10, r10, r9,  lsl #24
  12.393 +        ldr             r5,  [r2, #8]
  12.394 +        eor             r11, r8,  r10
  12.395 +        uhadd8          r7,  r8,  r10
  12.396 +        orr             r6,  r6,  r5,  lsl #24
  12.397 +        and             r11, r11, lr
  12.398 +        uadd8           r7,  r7,  r11
  12.399 +        ldr             r8,  [r1, #4]
  12.400 +        usada8          r0,  r4,  r7,  r0
  12.401 +        eor             r7,  r9,  r6
  12.402 +        lsr             r10, r5,  #8
  12.403 +        and             r7,  r7,  lr
  12.404 +        uhadd8          r4,  r9,  r6
  12.405 +        ldr             r6,  [r2, #12]
  12.406 +        uadd8           r4,  r4,  r7
  12.407 +        pld             [r1, r3]
  12.408 +        orr             r10, r10, r6,  lsl #24
  12.409 +        usada8          r0,  r8,  r4,  r0
  12.410 +        ldr             r4,  [r1, #8]
  12.411 +        eor             r11, r5,  r10
  12.412 +        ldrb            r7,  [r2, #16]
  12.413 +        and             r11, r11, lr
  12.414 +        uhadd8          r8,  r5,  r10
  12.415 +        ldr             r5,  [r1, #12]
  12.416 +        uadd8           r8,  r8,  r11
  12.417 +        pld             [r2, r3]
  12.418 +        lsr             r10, r6,  #8
  12.419 +        usada8          r0,  r4,  r8,  r0
  12.420 +        orr             r10, r10, r7,  lsl #24
  12.421 +        subs            r12,  r12,  #1
  12.422 +        eor             r11, r6,  r10
  12.423 +        add             r1,  r1,  r3
  12.424 +        uhadd8          r9,  r6,  r10
  12.425 +        and             r11, r11, lr
  12.426 +        uadd8           r9,  r9,  r11
  12.427 +        add             r2,  r2,  r3
  12.428 +        usada8          r0,  r5,  r9,  r0
  12.429 +        bgt             1b
  12.430 +
  12.431 +        pop             {r4-r11, pc}
  12.432 +endfunc
  12.433 +
  12.434 +.macro  usad_y2         p0,  p1,  p2,  p3,  n0,  n1,  n2,  n3
  12.435 +        ldr             \n0, [r2]
  12.436 +        eor             \n1, \p0, \n0
  12.437 +        uhadd8          \p0, \p0, \n0
  12.438 +        and             \n1, \n1, lr
  12.439 +        ldr             \n2, [r1]
  12.440 +        uadd8           \p0, \p0, \n1
  12.441 +        ldr             \n1, [r2, #4]
  12.442 +        usada8          r0,  \p0, \n2, r0
  12.443 +        pld             [r1,  r3]
  12.444 +        eor             \n3, \p1, \n1
  12.445 +        uhadd8          \p1, \p1, \n1
  12.446 +        and             \n3, \n3, lr
  12.447 +        ldr             \p0, [r1, #4]
  12.448 +        uadd8           \p1, \p1, \n3
  12.449 +        ldr             \n2, [r2, #8]
  12.450 +        usada8          r0,  \p1, \p0, r0
  12.451 +        pld             [r2,  r3]
  12.452 +        eor             \p0, \p2, \n2
  12.453 +        uhadd8          \p2, \p2, \n2
  12.454 +        and             \p0, \p0, lr
  12.455 +        ldr             \p1, [r1, #8]
  12.456 +        uadd8           \p2, \p2, \p0
  12.457 +        ldr             \n3, [r2, #12]
  12.458 +        usada8          r0,  \p2, \p1, r0
  12.459 +        eor             \p1, \p3, \n3
  12.460 +        uhadd8          \p3, \p3, \n3
  12.461 +        and             \p1, \p1, lr
  12.462 +        ldr             \p0,  [r1, #12]
  12.463 +        uadd8           \p3, \p3, \p1
  12.464 +        add             r1,  r1,  r3
  12.465 +        usada8          r0,  \p3, \p0,  r0
  12.466 +        add             r2,  r2,  r3
  12.467 +.endm
  12.468 +
  12.469 +function ff_pix_abs16_y2_armv6, export=1
  12.470 +        pld             [r1]
  12.471 +        pld             [r2]
  12.472 +        ldr             r12, [sp]
  12.473 +        push            {r4-r11, lr}
  12.474 +        mov             r0,  #0
  12.475 +        mov             lr,  #1
  12.476 +        orr             lr,  lr,  lr,  lsl #8
  12.477 +        orr             lr,  lr,  lr,  lsl #16
  12.478 +        ldr             r4,  [r2]
  12.479 +        ldr             r5,  [r2, #4]
  12.480 +        ldr             r6,  [r2, #8]
  12.481 +        ldr             r7,  [r2, #12]
  12.482 +        add             r2,  r2,  r3
  12.483 +1:
  12.484 +        usad_y2         r4,  r5,  r6,  r7,  r8,  r9,  r10, r11
  12.485 +        subs            r12, r12, #2
  12.486 +        usad_y2         r8,  r9,  r10, r11, r4,  r5,  r6,  r7
  12.487 +        bgt             1b
  12.488 +
  12.489 +        pop             {r4-r11, pc}
  12.490 +endfunc
  12.491 +
  12.492 +function ff_pix_abs8_armv6, export=1
  12.493 +        pld             [r2, r3]
  12.494 +        ldr             r12, [sp]
  12.495 +        push            {r4-r9, lr}
  12.496 +        mov             r0,  #0
  12.497 +        mov             lr,  #0
  12.498 +        ldrd            r4,  r5,  [r1], r3
  12.499 +1:
  12.500 +        subs            r12, r12, #2
  12.501 +        ldr             r7,  [r2, #4]
  12.502 +        ldr             r6,  [r2], r3
  12.503 +        ldrd            r8,  r9,  [r1], r3
  12.504 +        usada8          r0,  r4,  r6,  r0
  12.505 +        pld             [r2, r3]
  12.506 +        usada8          lr,  r5,  r7,  lr
  12.507 +        ldr             r7,  [r2, #4]
  12.508 +        ldr             r6,  [r2], r3
  12.509 +        beq             2f
  12.510 +        ldrd            r4,  r5,  [r1], r3
  12.511 +        usada8          r0,  r8,  r6,  r0
  12.512 +        pld             [r2, r3]
  12.513 +        usada8          lr,  r9,  r7,  lr
  12.514 +        b               1b
  12.515 +2:
  12.516 +        usada8          r0,  r8,  r6,  r0
  12.517 +        usada8          lr,  r9,  r7,  lr
  12.518 +        add             r0,  r0,  lr
  12.519 +        pop             {r4-r9, pc}
  12.520 +endfunc
  12.521 +
  12.522 +function ff_sse16_armv6, export=1
  12.523 +        ldr             r12, [sp]
  12.524 +        push            {r4-r9, lr}
  12.525 +        mov             r0,  #0
  12.526 +1:
  12.527 +        ldrd            r4,  r5,  [r1]
  12.528 +        ldr             r8,  [r2]
  12.529 +        uxtb16          lr,  r4
  12.530 +        uxtb16          r4,  r4,  ror #8
  12.531 +        uxtb16          r9,  r8
  12.532 +        uxtb16          r8,  r8,  ror #8
  12.533 +        ldr             r7,  [r2, #4]
  12.534 +        usub16          lr,  lr,  r9
  12.535 +        usub16          r4,  r4,  r8
  12.536 +        smlad           r0,  lr,  lr,  r0
  12.537 +        uxtb16          r6,  r5
  12.538 +        uxtb16          lr,  r5,  ror #8
  12.539 +        uxtb16          r8,  r7
  12.540 +        uxtb16          r9,  r7,  ror #8
  12.541 +        smlad           r0,  r4,  r4,  r0
  12.542 +        ldrd            r4,  r5,  [r1, #8]
  12.543 +        usub16          r6,  r6,  r8
  12.544 +        usub16          r8,  lr,  r9
  12.545 +        ldr             r7,  [r2, #8]
  12.546 +        smlad           r0,  r6,  r6,  r0
  12.547 +        uxtb16          lr,  r4
  12.548 +        uxtb16          r4,  r4,  ror #8
  12.549 +        uxtb16          r9,  r7
  12.550 +        uxtb16          r7,  r7, ror #8
  12.551 +        smlad           r0,  r8,  r8,  r0
  12.552 +        ldr             r8,  [r2, #12]
  12.553 +        usub16          lr,  lr,  r9
  12.554 +        usub16          r4,  r4,  r7
  12.555 +        smlad           r0,  lr,  lr,  r0
  12.556 +        uxtb16          r6,  r5
  12.557 +        uxtb16          r5,  r5,  ror #8
  12.558 +        uxtb16          r9,  r8
  12.559 +        uxtb16          r8,  r8,  ror #8
  12.560 +        smlad           r0,  r4,  r4,  r0
  12.561 +        usub16          r6,  r6,  r9
  12.562 +        usub16          r5,  r5,  r8
  12.563 +        smlad           r0,  r6,  r6,  r0
  12.564 +        add             r1,  r1,  r3
  12.565 +        add             r2,  r2,  r3
  12.566 +        subs            r12, r12, #1
  12.567 +        smlad           r0,  r5,  r5,  r0
  12.568 +        bgt             1b
  12.569 +
  12.570 +        pop             {r4-r9, pc}
  12.571 +endfunc
  12.572 +
  12.573 +function ff_pix_norm1_armv6, export=1
  12.574 +        push            {r4-r6, lr}
  12.575 +        mov             r12, #16
  12.576 +        mov             lr,  #0
  12.577 +1:
  12.578 +        ldm             r0,  {r2-r5}
  12.579 +        uxtb16          r6,  r2
  12.580 +        uxtb16          r2,  r2,  ror #8
  12.581 +        smlad           lr,  r6,  r6,  lr
  12.582 +        uxtb16          r6,  r3
  12.583 +        smlad           lr,  r2,  r2,  lr
  12.584 +        uxtb16          r3,  r3,  ror #8
  12.585 +        smlad           lr,  r6,  r6,  lr
  12.586 +        uxtb16          r6,  r4
  12.587 +        smlad           lr,  r3,  r3,  lr
  12.588 +        uxtb16          r4,  r4,  ror #8
  12.589 +        smlad           lr,  r6,  r6,  lr
  12.590 +        uxtb16          r6,  r5
  12.591 +        smlad           lr,  r4,  r4,  lr
  12.592 +        uxtb16          r5,  r5,  ror #8
  12.593 +        smlad           lr,  r6,  r6,  lr
  12.594 +        subs            r12, r12, #1
  12.595 +        add             r0,  r0,  r1
  12.596 +        smlad           lr,  r5,  r5,  lr
  12.597 +        bgt             1b
  12.598 +
  12.599 +        mov             r0,  lr
  12.600 +        pop             {r4-r6, pc}
  12.601 +endfunc
  12.602 +
  12.603 +function ff_pix_sum_armv6, export=1
  12.604 +        push            {r4-r7, lr}
  12.605 +        mov             r12, #16
  12.606 +        mov             r2,  #0
  12.607 +        mov             r3,  #0
  12.608 +        mov             lr,  #0
  12.609 +        ldr             r4,  [r0]
  12.610 +1:
  12.611 +        subs            r12, r12, #1
  12.612 +        ldr             r5,  [r0, #4]
  12.613 +        usada8          r2,  r4,  lr,  r2
  12.614 +        ldr             r6,  [r0, #8]
  12.615 +        usada8          r3,  r5,  lr,  r3
  12.616 +        ldr             r7,  [r0, #12]
  12.617 +        usada8          r2,  r6,  lr,  r2
  12.618 +        beq             2f
  12.619 +        ldr             r4,  [r0, r1]!
  12.620 +        usada8          r3,  r7,  lr,  r3
  12.621 +        bgt             1b
  12.622 +2:
  12.623 +        usada8          r3,  r7,  lr,  r3
  12.624 +        add             r0,  r2,  r3
  12.625 +        pop             {r4-r7, pc}
  12.626 +endfunc

    13.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
    13.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dsputil_init_arm.c	Mon Aug 27 12:09:56 2012 +0200
    13.3 @@ -0,0 +1,112 @@
    13.4 +/*
    13.5 + * ARM optimized DSP utils
    13.6 + * Copyright (c) 2001 Lionel Ulmer
    13.7 + *
    13.8 + * This file is part of FFmpeg.
    13.9 + *
   13.10 + * FFmpeg is free software; you can redistribute it and/or
   13.11 + * modify it under the terms of the GNU Lesser General Public
   13.12 + * License as published by the Free Software Foundation; either
   13.13 + * version 2.1 of the License, or (at your option) any later version.
   13.14 + *
   13.15 + * FFmpeg is distributed in the hope that it will be useful,
   13.16 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
   13.17 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   13.18 + * Lesser General Public License for more details.
   13.19 + *
   13.20 + * You should have received a copy of the GNU Lesser General Public
   13.21 + * License along with FFmpeg; if not, write to the Free Software
   13.22 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
   13.23 + */
   13.24 +
   13.25 +#include "libavcodec/dsputil.h"
   13.26 +#include "dsputil_arm.h"
   13.27 +
   13.28 +void ff_j_rev_dct_arm(DCTELEM *data);
   13.29 +void ff_simple_idct_arm(DCTELEM *data);
   13.30 +
   13.31 +/* XXX: local hack */
   13.32 +static void (*ff_put_pixels_clamped)(const DCTELEM *block, uint8_t *pixels, int line_size);
   13.33 +static void (*ff_add_pixels_clamped)(const DCTELEM *block, uint8_t *pixels, int line_size);
   13.34 +
   13.35 +void ff_put_pixels8_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.36 +void ff_put_pixels8_x2_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.37 +void ff_put_pixels8_y2_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.38 +void ff_put_pixels8_xy2_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.39 +
   13.40 +void ff_put_no_rnd_pixels8_x2_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.41 +void ff_put_no_rnd_pixels8_y2_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.42 +void ff_put_no_rnd_pixels8_xy2_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.43 +
   13.44 +void ff_put_pixels16_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h);
   13.45 +
   13.46 +CALL_2X_PIXELS(ff_put_pixels16_x2_arm,         ff_put_pixels8_x2_arm,        8)
   13.47 +CALL_2X_PIXELS(ff_put_pixels16_y2_arm,         ff_put_pixels8_y2_arm,        8)
   13.48 +CALL_2X_PIXELS(ff_put_pixels16_xy2_arm,        ff_put_pixels8_xy2_arm,       8)
   13.49 +CALL_2X_PIXELS(ff_put_no_rnd_pixels16_x2_arm,  ff_put_no_rnd_pixels8_x2_arm, 8)
   13.50 +CALL_2X_PIXELS(ff_put_no_rnd_pixels16_y2_arm,  ff_put_no_rnd_pixels8_y2_arm, 8)
   13.51 +CALL_2X_PIXELS(ff_put_no_rnd_pixels16_xy2_arm, ff_put_no_rnd_pixels8_xy2_arm,8)
   13.52 +
   13.53 +void ff_add_pixels_clamped_arm(const DCTELEM *block, uint8_t *dest,
   13.54 +                               int line_size);
   13.55 +
   13.56 +/* XXX: those functions should be suppressed ASAP when all IDCTs are
   13.57 +   converted */
   13.58 +static void j_rev_dct_arm_put(uint8_t *dest, int line_size, DCTELEM *block)
   13.59 +{
   13.60 +    ff_j_rev_dct_arm (block);
   13.61 +    ff_put_pixels_clamped(block, dest, line_size);
   13.62 +}
   13.63 +static void j_rev_dct_arm_add(uint8_t *dest, int line_size, DCTELEM *block)
   13.64 +{
   13.65 +    ff_j_rev_dct_arm (block);
   13.66 +    ff_add_pixels_clamped(block, dest, line_size);
   13.67 +}
   13.68 +static void simple_idct_arm_put(uint8_t *dest, int line_size, DCTELEM *block)
   13.69 +{
   13.70 +    ff_simple_idct_arm (block);
   13.71 +    ff_put_pixels_clamped(block, dest, line_size);
   13.72 +}
   13.73 +static void simple_idct_arm_add(uint8_t *dest, int line_size, DCTELEM *block)
   13.74 +{
   13.75 +    ff_simple_idct_arm (block);
   13.76 +    ff_add_pixels_clamped(block, dest, line_size);
   13.77 +}
   13.78 +
   13.79 +int mm_support(void)
   13.80 +{
   13.81 +    return HAVE_IWMMXT * FF_MM_IWMMXT;
   13.82 +}
   13.83 +
   13.84 +void dsputil_init_arm(DSPContext* c)
   13.85 +{
   13.86 +    ff_put_pixels_clamped = c->put_pixels_clamped;
   13.87 +    ff_add_pixels_clamped = c->add_pixels_clamped;
   13.88 +  
   13.89 +    c->idct_put              = simple_idct_arm_put;
   13.90 +    c->idct_add              = simple_idct_arm_add;
   13.91 +    c->idct                  = ff_simple_idct_arm;
   13.92 +    c->idct_permutation_type = FF_NO_IDCT_PERM;
   13.93 +
   13.94 +    c->add_pixels_clamped = ff_add_pixels_clamped_arm;
   13.95 +
   13.96 +    c->put_pixels_tab[0][0] = ff_put_pixels16_arm;
   13.97 +    c->put_pixels_tab[0][1] = ff_put_pixels16_x2_arm;
   13.98 +    c->put_pixels_tab[0][2] = ff_put_pixels16_y2_arm;
   13.99 +    c->put_pixels_tab[0][3] = ff_put_pixels16_xy2_arm;
  13.100 +    c->put_pixels_tab[1][0] = ff_put_pixels8_arm;
  13.101 +    c->put_pixels_tab[1][1] = ff_put_pixels8_x2_arm;
  13.102 +    c->put_pixels_tab[1][2] = ff_put_pixels8_y2_arm;
  13.103 +    c->put_pixels_tab[1][3] = ff_put_pixels8_xy2_arm;
  13.104 +
  13.105 +    c->put_no_rnd_pixels_tab[0][0] = ff_put_pixels16_arm;
  13.106 +    c->put_no_rnd_pixels_tab[0][1] = ff_put_no_rnd_pixels16_x2_arm;
  13.107 +    c->put_no_rnd_pixels_tab[0][2] = ff_put_no_rnd_pixels16_y2_arm;
  13.108 +    c->put_no_rnd_pixels_tab[0][3] = ff_put_no_rnd_pixels16_xy2_arm;
  13.109 +    c->put_no_rnd_pixels_tab[1][0] = ff_put_pixels8_arm;
  13.110 +    c->put_no_rnd_pixels_tab[1][1] = ff_put_no_rnd_pixels8_x2_arm;
  13.111 +    c->put_no_rnd_pixels_tab[1][2] = ff_put_no_rnd_pixels8_y2_arm;
  13.112 +    c->put_no_rnd_pixels_tab[1][3] = ff_put_no_rnd_pixels8_xy2_arm;
  13.113 +
  13.114 +    if (HAVE_NEON)    ff_dsputil_init_neon(c);
  13.115 +}

    14.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
    14.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dsputil_init_armv5te.c	Mon Aug 27 12:09:56 2012 +0200
    14.3 @@ -0,0 +1,41 @@
    14.4 +/*
    14.5 + * Copyright (c) 2009 Mans Rullgard <mans@mansr.com>
    14.6 + *
    14.7 + * This file is part of FFmpeg.
    14.8 + *
    14.9 + * FFmpeg is free software; you can redistribute it and/or
   14.10 + * modify it under the terms of the GNU Lesser General Public
   14.11 + * License as published by the Free Software Foundation; either
   14.12 + * version 2.1 of the License, or (at your option) any later version.
   14.13 + *
   14.14 + * FFmpeg is distributed in the hope that it will be useful,
   14.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
   14.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   14.17 + * Lesser General Public License for more details.
   14.18 + *
   14.19 + * You should have received a copy of the GNU Lesser General Public
   14.20 + * License along with FFmpeg; if not, write to the Free Software
   14.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
   14.22 + */
   14.23 +
   14.24 +#include "libavcodec/dsputil.h"
   14.25 +#include "dsputil_arm.h"
   14.26 +
   14.27 +void ff_simple_idct_armv5te(DCTELEM *data);
   14.28 +void ff_simple_idct_put_armv5te(uint8_t *dest, int line_size, DCTELEM *data);
   14.29 +void ff_simple_idct_add_armv5te(uint8_t *dest, int line_size, DCTELEM *data);
   14.30 +
   14.31 +void ff_prefetch_arm(void *mem, int stride, int h);
   14.32 +
   14.33 +void av_cold ff_dsputil_init_armv5te(DSPContext* c, AVCodecContext *avctx)
   14.34 +{
   14.35 +    if (!avctx->lowres && (avctx->idct_algo == FF_IDCT_AUTO ||
   14.36 +                           avctx->idct_algo == FF_IDCT_SIMPLEARMV5TE)) {
   14.37 +        c->idct_put              = ff_simple_idct_put_armv5te;
   14.38 +        c->idct_add              = ff_simple_idct_add_armv5te;
   14.39 +        c->idct                  = ff_simple_idct_armv5te;
   14.40 +        c->idct_permutation_type = FF_NO_IDCT_PERM;
   14.41 +    }
   14.42 +
   14.43 +    c->prefetch = ff_prefetch_arm;
   14.44 +}

    15.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
    15.2 +++ b/ffmpeg_smp/h264dec/libavcodec/arm/dsputil_init_armv6.c	Mon Aug 27 12:09:56 2012 +0200
    15.3 @@ -0,0 +1,121 @@
    15.4 +/*
    15.5 + * Copyright (c) 2009 Mans Rullgard <mans@mansr.com>
    15.6 + *
    15.7 + * This file is part of FFmpeg.
    15.8 + *
    15.9 + * FFmpeg is free software; you can redistribute it and/or
   15.10 + * modify it under the terms of the GNU Lesser General Public
   15.11 + * License as published by the Free Software Foundation; either
   15.12 + * version 2.1 of the License, or (at your option) any later version.
   15.13 + *
   15.14 + * FFmpeg is distributed in the hope that it will be useful,
   15.15 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
   15.16 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   15.17 + * Lesser General Public License for more details.
   15.18 + *
   15.19 + * You should have received a copy of the GNU Lesser General Public
   15.20 + * License along with FFmpeg; if not, write to the Free Software
   15.21 + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
   15.22 + */
   15.23 +
   15.24 +#include <stdint.h>
   15.25 +
   15.26 +#include "libavcodec/avcodec.h"
   15.27 +#include "libavcodec/dsputil.h"
   15.28 +#include "dsputil_arm.h"
   15.29 +
   15.30 +void ff_simple_idct_armv6(DCTELEM *data);
   15.31 +void ff_simple_idct_put_armv6(uint8_t *dest, int line_size, DCTELEM *data);
   15.32 +void ff_simple_idct_add_armv6(uint8_t *dest, int line_size, DCTELEM *data);
   15.33 +
   15.34 +void ff_put_pixels16_armv6(uint8_t *, const uint8_t *, int, int);
   15.35 +void ff_put_pixels16_x2_armv6(uint8_t *, const uint8_t *, int, int);
   15.36 +void ff_put_pixels16_y2_armv6(uint8_t *, const uint8_t *, int, int);
   15.37 +
   15.38 +void ff_put_pixels16_x2_no_rnd_armv6(uint8_t *, const uint8_t *, int, int);
   15.39 +void ff_put_pixels16_y2_no_rnd_armv6(uint8_t *, const uint8_t *, int, int);
   15.40 +
   15.41 +void ff_avg_pixels16_armv6(uint8_t *, const uint8_t *, int, int);
   15.42 +
   15.43 +void ff_put_pixels8_armv6(uint8_t *, const uint8_t *, int, int);
   15.44 +void ff_put_pixels8_x2_armv6(uint8_t *, const uint8_t *, int, int);
   15.45 +void ff_put_pixels8_y2_armv6(uint8_t *, const uint8_t *, int, int);
   15.46 +
   15.47 +void ff_put_pixels8_x2_no_rnd_armv6(uint8_t *, const uint8_t *, int, int);
   15.48 +void ff_put_pixels8_y2_no_rnd_armv6(uint8_t *, const uint8_t *, int, int);
   15.49 +
   15.50 +void ff_avg_pixels8_armv6(uint8_t *, const uint8_t *, int, int);
   15.51 +
   15.52 +void ff_add_pixels_clamped_armv6(const DCTELEM *block,
   15.53 +                                 uint8_t *restrict pixels,
   15.54 +                                 int line_size);
   15.55 +
   15.56 +void ff_get_pixels_armv6(DCTELEM *block, const uint8_t *pixels, int stride);
   15.57 +void ff_diff_pixels_armv6(DCTELEM *block, const uint8_t *s1,
   15.58 +                          const uint8_t *s2, int stride);
   15.59 +
   15.60 +int ff_pix_abs16_armv6(void *s, uint8_t *blk1, uint8_t *blk2,
   15.61 +                       int line_size, int h);
   15.62 +int ff_pix_abs16_x2_armv6(void *s, uint8_t *blk1, uint8_t *blk2,
   15.63 +                          int line_size, int h);
   15.64 +int ff_pix_abs16_y2_armv6(void *s, uint8_t *blk1, uint8_t *blk2,
   15.65 +                          int line_size, int h);
   15.66 +
   15.67 +int ff_pix_abs8_armv6(void *s, uint8_t *blk1, uint8_t *blk2,
   15.68 +                       int line_size, int h);
   15.69 +
   15.70 +int ff_sse16_armv6(void *s, uint8_t *blk1, uint8_t *blk2,
   15.71 +                   int line_size, int h);
   15.72 +
   15.73 +int ff_pix_norm1_armv6(uint8_t *pix, int line_size);
   15.74 +int ff_pix_sum_armv6(uint8_t *pix, int line_size);
   15.75 +
   15.76 +void av_cold ff_dsputil_init_armv6(DSPContext* c, AVCodecContext *avctx)
   15.77 +{
   15.78 +    if (!avctx->lowres && (avctx->idct_algo == FF_IDCT_AUTO ||
   15.79 +                           avctx->idct_algo == FF_IDCT_SIMPLEARMV6)) {
   15.80 +        c->idct_put              = ff_simple_idct_put_armv6;
   15.81 +        c->idct_add              = ff_simple_idct_add_armv6;
   15.82 +        c->idct                  = ff_simple_idct_armv6;
   15.83 +        c->idct_permutation_type = FF_LIBMPEG2_IDCT_PERM;
   15.84 +    }
   15.85 +
   15.86 +    c->put_pixels_tab[0][0] = ff_put_pixels16_armv6;
   15.87 +    c->put_pixels_tab[0][1] = ff_put_pixels16_x2_armv6;
   15.88 +    c->put_pixels_tab[0][2] = ff_put_pixels16_y2_armv6;
   15.89 +/*     c->put_pixels_tab[0][3] = ff_put_pixels16_xy2_armv6; */
   15.90 +    c->put_pixels_tab[1][0] = ff_put_pixels8_armv6;
   15.91 +    c->put_pixels_tab[1][1] = ff_put_pixels8_x2_armv6;
   15.92 +    c->put_pixels_tab[1][2] = ff_put_pixels8_y2_armv6;
   15.93 +/*     c->put_pixels_tab[1][3] = ff_put_pixels8_xy2_armv6; */
   15.94 +
   15.95 +    c->put_no_rnd_pixels_tab[0][0] = ff_put_pixels16_armv6;
   15.96 +    c->put_no_rnd_pixels_tab[0][1] = ff_put_pixels16_x2_no_rnd_armv6;
   15.97 +    c->put_no_rnd_pixels_tab[0][2] = ff_put_pixels16_y2_no_rnd_armv6;
   15.98 +/*     c->put_no_rnd_pixels_tab[0][3] = ff_put_pixels16_xy2_no_rnd_armv6; */
   15.99 +    c->put_no_rnd_pixels_tab[1][0] = ff_put_pixels8_armv6;
  15.100 +    c->put_no_rnd_pixels_tab[1][1] = ff_put_pixels8_x2_no_rnd_armv6;
  15.101 +    c->put_no_rnd_pixels_tab[1][2] = ff_put_pixels8_y2_no_rnd_armv6;
  15.102 +/*     c->put_no_rnd_pixels_tab[1][3] = ff_put_pixels8_xy2_no_rnd_armv6; */
  15.103 +
  15.104 +    c->avg_pixels_tab[0][0] = ff_avg_pixels16_armv6;
  15.105 +    c->avg_pixels_tab[1][0] = ff_avg_pixels8_armv6;
  15.106 +
  15.107 +    c->add_pixels_clamped = ff_add_pixels_clamped_armv6;
  15.108 +    c->get_pixels = ff_get_pixels_armv6;
  15.109 +    c->diff_pixels = ff_diff_pixels_armv6;
  15.110 +
  15.111 +    c->pix_abs[0][0] = ff_pix_abs16_armv6;
  15.112 +    c->pix_abs[0][1] = ff_pix_abs16_x2_armv6;
  15.113 +    c->pix_abs[0][2] = ff_pix_abs16_y2_armv6;
  15.114 +
  15.115 +    c->pix_abs[1][0] = ff_pix_abs8_armv6;
  15.116 +
  15.117 +    c->sad[0] = ff_pix_abs16_armv6;
  15.118 +    c->sad[1] = ff_pix_abs8_armv6;
  15.119 +
  15.120 +    c->sse[0] = ff_sse16_armv6;
  15.121 +
  15.122 +    c->pix_norm1 = ff_pix_norm1_armv6;
  15.123 +    c->pix_sum   = ff_pix_sum_armv6;
  15.124 +}